Elsevier has recently put out a new statement giving details of some changes it has made. In their own words,
In February, we informed you of a series of important changes that we are making to how the Elsevier mathematics program will be run. In this letter, we would like to update you on where we currently stand, and inform you of some new initiatives we have undertaken based upon the feedback we have received from the community.
I have known for some time that they were going to make an announcement of this kind, and that it would involve something called “more flexible subject collections”. During that time I have become clearer in my mind what it is that I don’t like about bundling. So before Elsevier’s announcement, I had in mind some tests that I would apply, to see whether having these new collections would mitigate the problems with bundling. (Spoiler: they don’t.)
Imagine you’re in charge of a university library and you have a limited budget. How do you spend that budget? In a system without any form of bundling, if you had an extra chunk of money, you would look around for the best additional utility you could buy with that money — where utility would be something like the number of new journal pages you could buy, multiplied by the average benefit that each journal page contributes to your university. In addition, the following would apply.
(i) If a journal’s quality goes down or its price goes up, then you have the option of subscribing to a different one that is better value for money.
(ii) If your budget goes down, you can decide which journals you value least (weighted by price) and cancel subscriptions to those.
(iii) If somebody wants to set up a new journal that’s better than existing journals in some way (for example it might be cheaper, or of a higher standard, or both), then they have a chance of persuading you to subscribe to their new journal, which you can pay for by cancelling subscriptions to less good journals.
Compare that with the system we have at the moment. What we actually have is this.
(i) If a (bundled) journal’s quality goes down, you can’t cancel your subscription to it. Your only option is to cancel the entire bundling agreement, which is a very drastic step to take.
(ii) If your budget goes down, then your options are to cancel a bundling agreement or to cancel subscriptions to journals that are not part of bundling agreements. The latter is much easier, so smaller independent journals are far more vulnerable.
(iii) If somebody wants to set up a new journal that’s better in some way than an existing journal, they will have great difficulty getting libraries to subscribe to it, because they cannot save money by cancelling subscriptions to bundled journals.
Somebody from Elsevier — I think it was Alicia Wise — defended bundling on the grounds that it protects more obscure journals that might otherwise struggle to find enough subscriptions. It’s possible that that’s the case, but it also protects bad journals, and that is a far stronger effect. If Elsevier were truly interested in protecting good journals in obscure subjects, then it could do what some academic publishers do with books, and use profits from some journals to subsidize others.
To summarize, bundling makes proper competition between journals impossible. That is what I object to about it, and I judge any move made by Elsevier by whether it makes things better in this respect. With that in mind, let’s look at some more of the Elsevier letter.
We already indicated that our target is for all of our core mathematics titles to be priced at or below US$11 per article (equivalent to 50-60 cents per normal typeset page), placing us below most University presses, some societies and all other commercial competitors. That will lead to a number of our titles seeing further and significant price reductions from their next volumes.
Further to this, and in response to feedback from the community for more flexibility around the packages and collections that we offer to libraries, we will take the added step of defining a smaller subject collection (around 15-20 journals) with our key core mathematics titles. The definitive list of journals will be determined shortly but will in any case include journals like Advances in Mathematics, Journal of Algebra, Journal of Number Theory, Journal of Functional Analysis, Journal of Combinatorial Theory A and B, and European Journal of Combinatorics, available with the discount levels offered on our subject collections.
To clarify what this means, let me briefly explain what the current system is, as I understand it. Elsevier has something called its “Freedom Collection” (I cannot help being reminded of the famous Freedom Fries that were popular in the US soon after the invasion of Iraq), which works roughly as follows: a library buys a number of journals at more or less the list prices — the ones it really wants — and then Elsevier offers all the rest, at a very heavy discount. The result of this is that the average price the library pays for its journals is much smaller than the list price. A typical sort of amount for a major library to pay to Elsevier in total is a bit over £1,000,000 per year, of which almost all goes on the “core collection” and only a tiny fraction on the rest. This page has a more detailed description of how it works, and includes the information that if you terminate your bundling contract, then you no longer have access to back issues of the journals outside the core collection.
In other words, everything is set up to make it as difficult as possible to leave a bundling arrangement once you’ve agreed to it.
In the light of all this, what difference does a reduction in list prices make? Very little indeed. It means that if you take the plunge and stop subscribing to the bundle, then it will be cheaper now than it was before. But the list prices are still high enough to make that an unpalatable option, especially given all those back issues you will lose access to. So reducing list prices (which remain very high, even if they are lower than those of many other publishers) does virtually nothing to address the drawbacks of bundling.
Ah, but what about the “response to feedback from the community for more flexibility around the packages and collections that we offer to libraries”? That sounds much more promising, since it adds a new option.
At first it looks ideal: instead of being under considerable pressure to buy a bundle with lots of second-rate mathematics journals, one can now go for the smaller subject collection with the “key core mathematics titles”. But who will want to do this? A library such as Cambridge’s or UCL’s currently subscribes to the Freedom Collection. Can it save money by replacing the mathematics part of that collection by the new smaller subject collection? No. I think there are smaller subject collections in other subjects as well. Can a library save money by subscribing to all those instead of to the Freedom Collection? No. (That last answer comes from a conversation that Urs Hartl recently had with Laura Hassink, Elsevier’s VP for strategy and journal services and one of the two signatories to Elsevier’s update — the other being David Clark. Later I’ll mention some other things that Urs found out.)
There is one kind of institution that could benefit from the new subject collections, and that is mathematics institutes, since they do not subscribe to journals across all the sciences. To that small extent, this move by Elsevier is welcome, but as an answer to the problems of bundling it is woefully inadequate.
In February, we made the archives of 14 core mathematics journals open, from four years after publication, back to 1995, the year when we started publishing digitally. We made more scholarly mathematics content freely available than has ever occurred before. We have now gone further and expanded the open archives back to 1995 for 43 journals in mathematics and related areas. For a full listing of Elsevier journals with an Open Archive, please see our information page.
This is of course good news, even if one might wish for it to go further still. In particular, it would be nice to have journals available from earlier than 1995, since in mathematics there are plenty of papers we want to consult that go back much further than that.
Why don’t they go further back? It’s true that there is a natural boundary in 1995, since, as they say, they started publishing digitally only then. However, they have digitized their archive from much earlier than that, and libraries can subscribe to this earlier part of the archive. So why not make that free? This is what Urs Hartl has to say about Laura Hassink’s response to that question: “She gave two reasons preventing this. First these issues have been bought by some institutions quite recently and it would be unfair to them to open the issues now. I said, I would know a solution for this and she herself guessed that one could solve the problem by refunding the recent buyers. As a response she mentioned the second reason, that there are 1900 issues from societies who had transferred the rights on these issues to Elsevier and want to earn royalties in return.”
I’ve got a partial solution to the second problem: at least make the journal issues available that are not affected by that problem. So a third excuse is needed …
The rest of Elsevier’s letter has less to do with the bundling and pricing issues so I won’t say much here. They have increased the length of time that some of their reviewers receive free access to their journals — a very minor concession since most of their reviewers will be at universities that subscribe to Science Direct, though it may be useful to some people. (They mention retired mathematicians, for example.) There is a paragraph about access for the developing world, which is welcome and important, but it’s not clear that anything substantial has changed. Finally there are three paragraphs of fluff entitled “Support to the mathematics community”.
I’ll finish by mentioning two particularly notable parts of Urs Hartl’s conversation with Laura Hassink. I hope the first in particular will demonstrate why so many people care about the current state of academic publishing and want big changes.
At the University of Münster, which is Urs Hartl’s university, Elsevier increased the price of its Freedom Collection (which is negotiated individually with universities) by 20% per year between 2011 and 2013. Here is the explanation from Urs: “Ms Hassink explained to me that in the contract with Elsevier a ‘committed spend’ of the University of Muenster is fixed. Since the institutes individually canceled titles before 2011 the ‘committed spend’ was not reached and this caused Elsevier to make up for it in the price increase for the Freedom Collection. I replied that this coincides with what various librarians and colleagues had told me. Namely, that no matter what they tried to do, the total amount they pay for Elsevier did not go down. Ms Hassink agreed that in this particular case with the Freedom Collection in Muenster the price increase had this intention.”
The second point is to do with the secrecy that surrounds Elsevier’s negotiations with universities. One of the reasons that Laura Hassink gave to Urs about this was that it would be impossible to offer generous discounts to developing countries if all contracts were disclosed. (I think the argument here is that if you admit to giving a discount to one university then they’re all going to want one.) Urs countered that “according to Florian Breuer’s computations the University of Stellenbosch, South-Africa pays roughly the same than the University of Muenster, Germany.” See also this this comment about the situation in India.
Interestingly, when I tackled David Clark about this issue (we met a few weeks ago) the reason he gave for non-disclosure was completely different. He claimed that there was a danger that disclosing the agreements would be in breach of competition laws, since it would allow the big publishers to coordinate their prices. If that explanation is correct, then the law is an ass. After all, competition laws are there to protect consumers against monopoly power, whereas (i) the lack of transparency of Elsevier hugely enhances their power and (ii) Elsevier effectively is a monopoly, since if they charge more than you like, you can’t get the same journals from a different source.
What about another big concession by Elsevier — the dropping of their support for the Research Works Act? Let me remind you of something they said when they announced this.
While we continue to oppose government mandates in this area, Elsevier is withdrawing support for the Research Work Act itself. We hope this will address some of the concerns expressed and help create a less heated and more productive climate for our ongoing discussions with research funders.
They are as good as their word. Horizon 2020 is the European Union’s “Framework Programme for Research and Innovation”. If you scroll down to page 13 of a long document they have put out, you will find the following sentence at the top of the page: “Further steps will be taken towards Open Access, to ensure that research results are available to those who need them.” At the bottom of the page, it says, “These measures will also focus on communicating the outcomes of research to policy makers, companies, innovators and other researchers, including by promoting Open Access.” I am told by a source that I trust that Elsevier is lobbying very hard to get all mention of open access removed from the Horizon 2020 documents. They have learnt from their mistakes over the Research Works Act — now they do their lobbying behind the scenes. Or at least so I have been led to believe: if any Elsevier representative would like to deny it, then please feel free to do so in the comments below and I will add a link here to the denial. Edit: Alicia Wise has provided a denial in this comment below.
I should say before finishing that this post is my personal reaction to the Elsevier statement: I don’t speak for anyone else, and I don’t guarantee to have got all my facts right. I welcome corrections and alternative views in the comments below.
At the time of writing, the number of signatories to the boycott at the Cost of Knowledge page has just passed 11,000. The number of mathematicians is creeping up to 2,000. It would be nice to pass that milestone soon, and keep on climbing. Nothing that Elsevier has said gives us any reason to end the boycott. They are behaving much as one would expect: offering minimal concessions that will look as good as possible while keeping their profits intact. I realize that asking them to deal with the objections to bundling and exposing their journals to genuine competition is making a demand they are most unlikely to accede to, since their huge profits are based on stifling this competition. So instead, we must press on with the more positive step of developing alternative models, something I shall report on in the near future.
Despite Elsevier’s strong dislike for open-access mandates, there does appear to be significant movement in precisely that direction. Three interesting recent news stories are an announcement by the Wellcome Trust that it would insist on open access for the research it funds, a statement by Harvard University Library that the current system is unsustainable, and an announcement by the British Government that it has plans to make all taxpayer-funded research available online. It is difficult to move to new systems, but external pressure of this kind will surely help.
(Update: The International Association of Scientific, Technical and Medical Publishers (STM) has put out a press release responding to the British government announcement just mentioned. It’s worth a read. It is clear that the most important word in the press release is “sustainable” and that it means “capable of sustaining the profits of the major publishers”.)