Elsevier’s open letter point by point, and some further arguments

I’ve had this post sitting around half written for a long time, so I’ve decided to post it without saying everything I wanted to say and have done with it. Even the things I do say are not as organized as I’d normally like.

In response to the recent boycott and the discussion it has generated, Elsevier has put out an open letter entitled A message to the research community: journal prices, discounts and access. It was more or less guaranteed that I would not be satisfied by what they had to say, since my view is that the entire system of commercial publishing of academic papers needs to be replaced, whereas Elsevier was almost certain to be arguing from within the current system. When a paradigm shift takes place, one does not expect the main players to remain the same.

Nevertheless, it seems only fair to make some attempt to respond to the points Elsevier has to make, rather than simply saying that there is nothing they can do, or at least nothing they can do that they would be likely to be prepared to do. So let me explain why I find their open letter unsatisfactory even on their own terms (that is, even if we make the assumption that what is needed is more like an adjustment to the current system than a replacement of it). The letter starts with some statements of the kind one might expect, about how their mission is to serve the research community and so on. The substantive part of their letter is then divided into bullet points, where they claim to correct the distortions that have been advanced. I’ll look at each of those in turn.

First, the cost of downloading an article has never been lower than it is today — on average one fifth of what it was just 10 years ago. As the effective price paid per journal accessed has decreased, the number of journals accessed has increased, and the usage of those journals has grown by over 20% per year. We have invested heavily in making our content more discoverable and more accessible to end-users and to enable the research community to develop innovative research applications. Perhaps it shouldn’t be a surprise that in a study published by the Publishing Research Consortium, which Elsevier’s research team help conduct, 93% of university researchers report that access to journal articles is “fairly easy or very easy.” The full report is available here.

An annoying aspect of the open letter is that Elsevier are not precise about what they mean. For example, what is meant by the cost of downloading an article? Does it mean the cost for a private individual who follows a link to an article, clicks “Download pdf” and is then asked for $30 or so? If so, did it really cost $150 10 years ago? That would have been insane. Or does it mean what you get if you divide the total amount paid by all the institutional subscribers round the world by the total number of downloads by researchers at those institutions? If it’s the latter (or something more like the latter), then is the fall in this average cost due to a lowering of prices or to an increase in the number of downloads, or to some mixture of the two? And if there has been an increase in the number of downloads, is that not explained by the fact that everybody does more online these days, coupled with the fact that looking at a paper online takes much less effort than trekking over to the library and finding it on a shelf?

What does it mean to say “We have invested heavily in making our content more discoverable and more accessible to end-users and to enable the research community to develop innovative research applications.”? One possible interpretation of “making our content more discoverable” is the very annoying phenomenon where an internet search for a paper takes you to an Elsevier page that suggests that all you have to do is click to download the paper, but then asks for money when you do. That kind of ease of discovery I could do without. The phrase “and accessible” reads to me like pure padding: either I can get hold of an article or I can’t. And where is the heavy investment needed to allow someone to look at an article. The article is there. And what are the innovative research applications? I simply don’t understand what they are saying here, and I suspect that if it means anything at all, then whatever it means is of little interest to mathematicians.

As for the report they refer to, it again is very hard to interpret without a lot more information. My experience with Elsevier papers is this. If an author has taken the trouble to put a paper on the arXiv or on his or her homepage, then access to that article is easy (no thanks to Elsevier); otherwise, it is virtually impossible. Probably if I made more effort, I would be able to work out how to get access to Elsevier articles via Cambridge University’s subscriptions to their journals. But if I’m working from home, then I know that that is a complicated process that involves setting up some kind of account with our computer services — I tried it once but didn’t follow through to the end. In practice, if an article is behind the Elsevier paywall, I don’t read that article. Does that make me an outlier? Am I one of the 7%? Is my situation even part of what the question covers? I haven’t tried recently, but maybe if I tried to open an Elsevier article from the computer in my departmental office I would have no trouble. But all that would say is that Elsevier didn’t have a completely stupid system for opening articles. [Edit: Since writing that paragraph I have discovered that I can open Elsevier papers with no trouble at all from my office, and with only a small amount of effort from home. But others may not be so fortunate.]

I would suggest a different wording for the question: “How satisfied are you with the level of access that Elsevier provides for its articles?” Then the answers might be more informative. (I recommend looking at the report, which has no discussion of methodology whatsoever, as far as I can see.)

Libraries are never forced to take “bundled” packages; they always have the option to purchase individual articles, subscribe to titles, or subscribe to sets of journals. Most choose large collections, however, because they get substantial volume discounts that offer more titles at a lower cost. And the additional titles they subscribe to are used by their researchers. In fact, on average approximately 40% of researchers’ usage is of journal titles that the library previously had not subscribed to.

The devil is in the detail here, and because of the confidentiality agreements that Elsevier insists on, it is very hard to find out the detail. So let me start by asking an obvious question: if the pricing policy is so reasonable, then why keep it confidential? Is it not healthy for academics to know how much their libraries are spending for their benefit, and what they are getting for the money, and what other deals would have been possible?

Let me try to explain what I object to about bundling. I want to be clear that I am not trying to prove that it is immoral. I just want to explain why I dislike it. That’s all I need to do to explain why if it is feasible to change to a different system, then I am strongly in support of doing so.

Let me draw an (imperfect — I’ll come to that) analogy. Suppose you go to your usual supermarket and find that their prices have suddenly become extraordinarily high. Of course, you walk out, but suppose you are intercepted by an employee of the supermarket and asked whether you are aware of their new “hamper deals” — you get a hamper that contains a wide variety of food items, and it works out far cheaper than buying all those items individually. You have a look in the hamper and you find that it does indeed contain pretty much everything on your normal shopping list, but it also contains a number of things you didn’t plan to buy, some of which you would be quite pleased to have and others of which you would probably end up throwing away. You note to yourself that the hamper is somewhat more expensive than what you have been used to spending at this supermarket, even if it is indeed much cheaper than buying the items individually. But you do get those extra items, so you decide, slightly reluctantly, to go for the hamper deal.

Next time you go back to the supermarket, nothing has changed, so you go for the hamper deal again, and this continues for a few weeks. But then the supermarket starts raising the price of the hamper deal. It’s still a lot cheaper than buying the items individually — not surprisingly as the individual prices are ridiculous — but it is clear that prices are going up faster than inflation (or even than food prices in general).

Why might you dislike this situation? It’s fairly obvious: by accepting the hamper deal you have lost your control over what you buy from the supermarket, and once you are in the habit of buying their hampers, they can raise prices in the knowledge that your other options — buy just a few very expensive items individually or stop using the supermarket altogether — are unattractive to you.

Or are they? Here is where the analogy is not perfect. If a supermarket were to start messing about with its customers like that, then those customers would have the option of using a different supermarket, and it probably wouldn’t take much for them to exercise that option. The big difference between supermarkets and academic publishers is that supermarkets are competing to sell the same goods, whereas academic publishers are selling disjoint sets of journals. In other words, academic publishers have a monopoly over their sets of journals. It’s as though you have one supermarket chain for vegetables, one for meat and ready meals, one for cooking ingredients, cereal, tea, coffee, bread, and so on. And you can’t get vegetables at even semi-reasonable prices without subscribing to a regular delivery of a fairly expensive vegetable box, which always contains more than you need. Under such circumstances, one might imagine growing one’s own.

Even this isn’t a perfect analogy, and what I should probably be talking about is institutional purchase of vegetables. If I found out, for example, that, in order to provide my lunch, Trinity College was having to pay for very expensive deliveries of vegetables, many of which were not needed … actually, even this isn’t a good analogy. Let’s suppose in addition that the fellows of Trinity had grown the vegetables in their back gardens and given them to the vegetable delivery company (who packaged them up quite nicely) and were then finding that Trinity College was paying through the nose for them. At that point, I’d be tempted to say that I’d be happy to bring my vegetables directly to the college, not packaged as nicely but free.

To make the analogy even better, one might add that different colleges paid different amounts for their vegetables, but that how much they paid was confidential — presumably so that rich colleges like Trinity would pay more than poorer colleges, or to put it more crudely, so that the vegetable delivery company could screw each college for as much as it could afford.

I was thinking of discussing bundling in terms of a mathematical model, but that’s something I’m not going to bother with any more. The main point can be summarized as follows: Elsevier has a monopoly over a large number of journals (because you can’t just substitute them with other ones) and it exploits that monopoly. It’s welcome to do that, and those who pay for their journals (whether directly or indirectly) are welcome to find it irritating and try to find ways of doing without the product over which they have a monopoly.

Just before I finish, I’d like to consider two arguments that were put forward by another Elsevier representative, Alicia Wise. The first was that bundling protects some journals that would not otherwise get enough subscriptions to survive. To that I have two responses. One is that at least some of those journals are ones that shouldn’t survive — such as the notorious Chaos, Solitons and Fractals. I think her implication is that Elsevier is protecting the interests of journals in obscure but very worthwhile areas, but is it really those rather than bad journals in perfectly well-represented areas? I would need further evidence.

Her second point was that Elsevier adds value in many ways, such as Science Direct, where they have put all their papers with lots of hyperlinks and so on. I looked into that by going, for the first time in my life, to Science Direct and looking at a randomly chosen paper, entitled “Paley graphs have Hamiltonian decompositions” and published in Discrete Maths. I’ll admit it was quite a nice site. There was a brief section called “highlights”, which wasn’t part of the paper but gave you some idea what to look out for. There were internal links to references and to numbered statements within the paper. There were also links to any papers in the references that were published in Elsevier journals. Discrete Mathematics, not being a top-notch journal, is full of references to itself, so if you happen to be part of, or interested in, one of those mathematical conversations and have access to Science Direct, then you would probably find this feature quite convenient. They also had a list of related papers, but I’m not sure how useful that is, as it appeared to be automatically generated. For example there were several papers listed about Cayley graphs. (Is that because “Cayley” rhymes with “Paley”? Surely not, so exactly how the list was generated is a bit of a mystery to me and an amusing exercise in reverse engineering.)

Anyhow, my response to all that is that if that is the main way that Elsevier (as opposed to the unpaid referees and editors) adds value, then what I would like to see is a system where we paid Elsevier for adding that value. Under such a system, a journal would be an autonomous organization, but if it wanted its papers to appear on Science Direct, so that they had all those links (the buzzword seems to be “metadata”), then they could pay for that service. Of course, Elsevier would have to set the price at a level that would be worth paying, given what the journal was getting. But then I’d feel confident that Elsevier was getting a fair price for the value it was adding, whereas now I don’t at all.

As I warned at the beginning, I haven’t covered all the points I intended to cover, but I think it’s time to reclaim this blog for non-Elsevier-related posts. That doesn’t mean that there won’t be further news about Elsevier. There will, and my interest in the subject is as strong as ever. But the change that many of us have hoped for for a long time is not going to happen overnight, so I’m going to write about other things as well. Meanwhile, I hope that if you agree with the objectives of the boycott, then you will think about what you can do to help — things like making sure that your papers are easily available online, encouraging others to do the same, not submitting to an expensive journal when a cheap one is equally suitable (which I know is not always), supporting new cheap journals and publishing methods, and so on.