Tuesday, December 12, 2006

Scale and scalability

Scholarly publishing is a pretty large-scale pursuit. The results of every serious research project outside the intramural confines of industrial R&D must be interpreted and published, or the research is deemed not to have taken place. Even a lot of industrial research is published – though some, mainly in the more obscure journals, purely for the purpose of ‘prophylactic disclosure’, in case an invention is not deemed worthy of patenting, yet if patented by anyone else, could become a ransom threat (an invention that has been disclosed can never be patented anymore, a trick also used by ‘open sourcerers’ to ensure that their code cannot be appropriated). But I digress. If the estimate of 25,000 journals is in the right ballpark, and each publishes 40 articles per year on average, about a million new articles is being added to the literature every year. It’s probably even more than that. And if the average rejection rate is 50%, these 25,000 journals actually process at least 2 million articles per year. A number of these articles will ‘cascade’ through the journal pecking order and finally be published somewhere, having been processed and peer-reviewed several times (let’s hope the resulting fine-layered publication hierarchy is worth such a waste in the system).

A veritable industry, this scholarly publishing. Good that there are professional, independent organisations that take on the drudge of all that work. Perish the thought that researchers would have to organise it all by themselves.

So everybody is thankful for what the publishers do? No. There is a problem: they want to be paid for what they do, and they don’t even do what we want them to do, which is to give everybody free, open access to whatever research they publish.

Two ‘solutions’ have been proposed. One that deals with the cost of publishing only; and one that deals with open access only, ignoring any issues of cost.

The solution that deals with cost only is the one that holds that the costs are too high and it’s all the fault of ‘commercial’ publishers. Instead, all scientific publishing should be done by not-for-profit scholarly societies. On the face of it, journals published by these NfPs (NfP journals don’t exist – just NfP publishers, who still wish to see their journals turn a surplus – the non-tax-payer's equivalent of, or euphemism for, profit) do seem to have lower subscription charges. Which is sort of easy, if one makes a subscription a compulsory part of membership. And levies page charges. And pays no tax. Many society publishers can offer relatively low-priced subscriptions, and still make revenues that on a per-article basis, are similar to what is being realised by commercial – I prefer to call them independent – publishers. Or more, which is, given their NfP status, kept in reserve or spent on good causes, of course.

It certainly works. The point, however, is that such publishing is not scalable. If it would just be the NfP status of society publishers, an increase of scale would have happened long ago. Nobody has ever tried to stop NfPs from cheap journal publishing. But it isn’t the NfP status that is the cause of lower prices; it is the fact that the membership yields much of the revenues needed to support the journals. And imagine 25,000 journals each sustained by enough society members to result in low subscription prices. It just doesn’t stack up. Otherwise we should have seen strong growth in NfPs. Interestingly, the contrary took place. Independent publishers started to flourish because societies couldn’t deal with the growing volume of articles and the increased international and interdisciplinary nature of science. Many still can’t – particularly the ones with relatively low membership or relatively voluminous journals – and testimony of that is that a growing number of societies are ‘outsourcing’ their journal publishing to independent publishers.

Cheap subscriptions do work for some journals, but cannot work for all, and it has precious little to do with the NfP status of the publisher.

The ‘solution’ that deals with open access only, ignoring any issues of cost, is of course what is known as ‘self-archiving’. Self-archiving assumes that librarians paying for subscriptions that are not necessary anymore, keep journals economically viable. Self-archiving, after all, is only meant to “fulfil the access-needs of would-be users who cannot afford access to the proprietary journal.” So here you are, librarians: even though it’s all freely available, if you can afford to take a subscription, please do. Charity is a good thing, of course, but not exactly the most robust foundation for an activity that is so much part of research and without which much of the academic world would be lost: recording research results in peer-reviewed journals.

I couldn’t have said it better than Stevan Harnad himself, on 10 December 2006, on the AMSCI Open Access Forum (not yet archived there as I'm writing this): “I, for one, have never doubted that [publishers making journal articles open after a short embargo] could cause cancellations. But anarchic author self-archiving, of each author's postprints, in each author's own IR, in uncertain proportions and at uncertain rates, are [sic] another story”. Precisely.

Which means that self-archiving is also not scalable. As long as there is only a small number of authors engaged in self-archiving, and it is done anarchically and unpredictably, it will work. Publishers will have little practical problem with it and librarians will not cancel subscriptions on that basis alone. Take the anarchy and unpredictability out of it, however – for instance via self-archiving mandates – and it would all be, to borrow Stevan’s phrase, another story.

Putting it another way: if self-archiving were to succeed, it would fail. Succeeding, after all, means sufficiently increasing in scale to provide open access to a meaningful proportion of the literature. Which would, of course, lead to cancellations. No publisher, be it an independent or NfP, could afford to allow authors to self-archive in such circumstances, and ‘green’ would fade out of existence.

It is possible, of course, that the chaos of self-archiving leads to a phase transition to a stable and truly scalable method, open access publishing, a.k.a. ‘gold’, but it seems a rather circuitous and acrimonious route to take. Why not stimulate ‘gold’ straight away, especially since a rapidly increasing number of publishers offer it? Why not lobby for an open-access-mandate, instead of for a self-archiving-mandate? It can’t be the money, though it would change the relative proportion of costs, between institutions, and that’s probably where the rub is. It's understood that it needs to be done, but “you first, sir!” A sur-place as it’s known in track cycling, where whoever moves first lessens his chances of winning.

The advantages of ‘gold’ are huge. Immediate open access to the research literature; costs move – up or down – with the research activity itself; a functional market, with fair price levels as a consequence; discouragement of spurious, speculative, or ‘ultra-light’ submissions; elimination of visibility as an element in perceived quality; and I’m probably forgetting to mention a few.

Jan Velterop


  1. Jan misunderstands what I said: Journals making all their own contents OA (gold), all at once, and all in one place, right now, is likely to cause cancellations. In contrast, authors making their own articles OA by self-archiving them in their own IRs (green), anarchically and distributedly, does not provide 100% of the contents of any journal, and its extent and growth rate is hard to ascertain. Hence that is unlikely to cause journal cancellations until the self-archiving of all articles in all journals is reliably at or approaching 100%. If/when that happens, journals can and will downsize to becoming peer-review service providers only, recovering their much reduced costs on the OA model that Jan favors. But journals are very unlikely to want to do that downsizing and conversion now, when there is no pressure to do it. And there is certainly no reason for researchers to sit waiting, while they keep losing access, usage and impact. Mandates will pressure researchers to self-archive, and eventually 100% self-archiving might pressure journals into downsizing and converting. Right now, however, journals are all making ends meet through subscriptions, whereas researchers keep losing about half their potential daily usage and impact, cumulatively. The immediate priority of research, researchers, their institutions and their funders is hence obvious, and it is certainly not for authors, their institutions or their funders to pay their journals' current asking price for making each article OA, over and above paying for subscriptions, so that publishers can have their cake and eat it too: It is to make their own articles OA, right now, by self-archiving them, and paying for peer review only if and when journals have downsized, cut costs, and there is no other sustainable way of recovering their costs. (By that time, of course, subscription cancellation savings will be available to pay those reduced costs up-front. Today they are not; and double-paying up front would be pure folly.)

  2. 'Downsizing', 'reduced cost', 'double-paying', is Stevan's agenda possibly not just 'open access'?

    To repeat what I've said several times before: self-publishing is what authors can do freely, if they so wish. They can provide open access to their articles to their hearts' content. Once they involve a publisher, though, they don't do that out of altruistic motives. They don't 'give' their articles to publishers. They come to ask for a 'label' that makes their article from a piece of text, perhaps interesting, but not recognised by the academic community, into a formally peer-reviewed and published article. It's not the publishers that compel them to do that.

    To ask publishers to 'label' their articles in such a way (making them valuable for career and funding prospects) and then subsequently to take that label and stick it on the self-published, free version of an article, if done without paying, is depriving the publisher of a means to earn his keep by providing those services. Is this an honest practice? Or does the fact that it hasn't done much damage yet to the publishers make it more honest than it otherwise would be?

    And what does Stevan have in mind when he thinks about 'downsizing'? Give up print, perhaps. Which represents only a small fraction of the current cost. Many publishers would only be too happy to abolish print. But don't delude yourself with the idea that having no print would make much difference to prices.

    Instead of downsizing, what we'll see is new entrants, with an already pre-downsized structure, free of – and deprived of – any legacy, such as Hindawi or BMC. Stevan's self-publishing mandates make it difficult for such new entrants to succeed on any scale needed to provide the full open access that the world – not just the academic world – needs. If a full, safe cure for a disease is possible, though not necessarily cheaper than proper lifelong symptom-management, is it better to go for cheap palliative care than for this full cure?

    Jan Velterop