Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
Let us assume that journals as we know them will continue to exist for a substantial time. I am now close to convinced that no major technical issue prevents non-profit groups from using the subscription model and/or the gold OA (pay-to-publish) model to deliver journals much more cheaply than is currently done, with no loss in quality. When I started explaining to my wife why this issue was so important, she asked why the market does not take care of it. What are the barriers to entry to prevent someone undercutting Elsevier?
It seems to me that the main reasons are: 1) excessive reliance on poorly defined and slow to change journal-level prestige measures for evaluation of researcher quality and 2) authors don’t directly receive price signals, and are not necessarily able to judge the value for money of a journal. In other words I agree with this post by John C. Baez: http://johncarlosbaez.wordpress.com/2012/02/29/8410/.
I would like to discuss whether there are any other reasons that should be considered, whether the above obstacles are indeed obstacles, and what we might do about it. Economists are particularly welcome.
I am trying to keep each post short, so here is the next one. One thing I find suboptimal is the way prestige of journals is measured and I would like to see it changed - perhaps the AMS and IMU could address this and push universities to adopt best practices.
I see two ways currently used. One is appeal to folklore - “it is commonly accepted that”. This means that the speaker has heard this from his/her social network and sees no strong reason why it isn’t true. For those who for geographic and other reasons aren’t near the top 10 universities in the USA, this is a problem. I want to push the idea that mathematics is a global field and not confined to the top 10 places in the USA. Those are all very strong institutions, but there is a lot more out there. Tim Gowers had a recent post that discussed journals in combinatorics. I found it amazing that he had never looked at EJC until then, and when he did, he revised his opinion of it upwardly by a considerable amount.
An alternative to folklore would be a broad-based survey of thousands of researchers to get their idea of which journals are in each level. The Australian ARC ERA journal and conference ranking project http://www.arc.gov.au/era/journal_list_dev.htm was one such attempt. It seemed fine to me, but now after huge effort has been discontinued. Classifying into 4 categories seem enough level of detail. One thing I find odd is people claiming to make very fine distinctions (“Inventiones is much better than Annals of Math” - type comments) which a) are surely wrong and b) are irrelevant to most mathematicians, who will never publish in either and may not read anything from either anyway. I guess
Another method is citation statistics such as Journal Impact Factor. The IMU has a report on citation statistics http://www.mathunion.org/fileadmin/IMU/Report/CitationStatistics.pdf which savages their naive use. However I doubt they will go away, and IMU report doesn’t give any other reasonable alternatives (on my cursory reading). For a fixed research field, a citation-based method might well be better than folklore. A PageRank-like system that doesn’t weight all citations equally may be better. I recently checked eigenfactor.org and found the AI score there to correlate strongly with my prejudices.
Of course I maintain that for evaluation of researchers, article-level citation metrics will be much more informative than journal-level ones. But if a library is deciding on subscriptions, perhaps it should look at journal scores.
Bundling creates enormous barriers to entry. It more or less guarantees that the fraction of a library’s budget devoted to Elsevier journals (or those of any publisher using the same strategy) will gradually increase. The contracts I’ve seen have total costs that increase by 5% per year over a five-year period, with only very limited ability to cancel journals (if it can be done at all). You don’t have to renew a bundle when it expires, but then you’ll be paying list price, which is typically increasing even more rapidly.
In order to accommodate this growth, libraries have to cut something, typically subscriptions from smaller publishers, and they have practically no funds available for new subscriptions. This makes it incredibly hard to start a new journal, much harder than it ought to be. Publishers like MSP, which offer better quality than Elsevier at a tiny fraction of the price (8 cents per page for the MSP Mathematics Package, vs. Elsevier’s new, reduced target of 50 to 60 cents per page), find it extremely difficult to get enough subscriptions to get by. Elsevier and its large competitors have locked up way too much library funding into huge, inflexible contracts, and everyone else is fighting like crazy over the portion remaining.
I actually think this is a much bigger barrier than any of the sociological or psychological effects. For example, the MSP journals have become quite prestigious (much more so than all but a couple of Elsevier’s journals), and they have great bibliometrics. However, prestige just doesn’t translate into subscriptions when budgets are tight, and bundling guarantees permanently tight budgets for all journals that don’t come from the largest publishers.
In order to have a reasonable scholarly publishing landscape, we’re going to have to end these pricing policies. With luck, we’ll be able to do it through community pressure, such as the boycott. If necessary, we may have to explore options like antitrust law. (For a 2004 analysis, see http://www.law.berkeley.edu/faculty/rubinfeldd/Profile/publications/Edlin-Rubinfeld%20ALJ%20Paper.pdf. Of course I don’t have any legal expertise, but I believe the situation has grown considerably worse since then, and an antitrust case is not an implausible end game here.)
This reminds me of the joke about an economist refusing to pick up a $20 bill from the street, on the grounds that an efficient market ensures that it couldn’t possibly actually exist.
But of course Henry’s explanation is pretty much it.
Thanks Henry - that clarified things a lot.
On bundling and what to do about it: this list may be useful http://www.earlham.edu/~peters/fos/lists.htm#actions. It is interesting how few institutions have done anything. I guess the professional societies should make some clear statements to their members.
Another question: I have a meeting with our top university librarian and some other committee members next month. Any advice for what I should ask/tell them? I will try to find out what we pay for serials, broken down as much as possible.
This has been a very interesting discussion to follow. I have not a lot to contribute to the main thread, but I have some conclusions from what you two have said. It looks as though it would be very hard to break the hold of the big publishers by trying to siphon off their main journals to independent publishers (or some such). Doing it piecemeal, as it currently is being done, produces no drop in cost to the library on the existing journals (and even if the old version of the journal does not persist, there’s still the back issues to take into consideration) and so actually shifting journals away from commercial publishers is possibly not a viable strategy. This is perhaps a bit depressing, but on the other hand makes me more eager to find solutions to the publishing issue that are, in some sense, orthogonal to journals so that they can exist alongside journals while we still have journals in the system.
And if we can end bundling, then the system will begin to sort itself out automatically, through genuine price competition.
Unfortunately, I think that Mike’s last paragraph (#5) scuppers this. The Elsevier boycott did not say:
I refuse to read or cite any article published in an Elsevier journal.
Moreover, the backlog issue is immense. I remember hearing a radio show about baked beans (this might not make sense for non-UK readers!). Branston (of pickle fame) were trying to get into the market and, of course, the big name in baked beans is Heinz. So the radio show did a taste test and the person doing the test said something of the nature of that the Branston one was really nice and they really liked it, but it just wasn’t Heinz baked beans and Heinz had so many memories and associations for them, that there was no way they were going to buy another. Elsevier and so on have all our archives, so it’s going to be very hard to compete with that.
I’ve been meaning to comment here…
If you were an investor back in Henry Ford’s day and you convinced yourself that one of these days, those crazy automobiles would fill the streets and change transportation forever, which of the many automobile companies should you invest in? Pick the wrong one and you’re bust. Choose too many and you don’t win because you’ve invested too little in the winners.
The answer: short horses.
One way to bring markets into this is to fund a team to develop the technology, lobby the universities, and market the products that would end Elsevier and other publishers. How do you fund it and where is the market incentive?
The answer: short publishers.
PS: I’m seriously considering this.
Hi Henry :)
Yeah, I meant short selling. Not tiny horses :)
Getting size is not an issue. The numbers in finance are mind numbing. It would not take more than a few 10s of million USD to fund the project and that is a neglible drop in the bucket. It is risky and there would need to be someone willing to back the project if the stock moved against you to avoid getting stopped out from a margin call, but it could be very profitable if the project was a success.
After all, fundamental research is government-funded anyway.
Only some of it, and a minority of mathematics research, even if you use a broad definition of government funding (to include things like tax deductions for donations to universities).
removing the mechanism by which they can overcharge and bundle
Actually, government contracts are part of the mechanism by which Elsevier can overcharge and bundle. They love to make deals at as high a level as possible (multiple campuses or even entire countries at once), perhaps partly for the sake of efficiency, but seemingly also because it takes control away from faculty and librarians. The nightmare scenario is that Elsevier and some government officials work out a deal and then impose it on everyone. That already happens too often, and we certainly shouldn’t encourage it.
Overall, I’d push for exactly the opposite approach, namely trying to minimize the role of governments in making any funding decisions for journals:
(1) They aren’t competent to do a good job in any way except by asking researchers and librarians what to do, and I don’t trust them to do that reliably in the long run. Even if they started with good intentions, we could eventually end up with a bureaucratic nightmare involving impact factors and other bibliometrics.
(2) There’s way too much potential for political influence. If the government is in charge, then that’s tantamount to putting lobbyists and interest groups in charge. Elsevier has more money and more lobbying experience than we do, and that translates into more political influence.
(3) We cannot count on consistent levels of government support: budgets fluctuate due to government priorities, the economy, etc., and we’ve seen disastrous effects on state universities. We have to deal with these fluctuations no matter what, but we should try to isolate journal funding from them as much as possible, since archiving and journals are intended to be permanent. The government may be able to cut math funding, but they should not be able to specifically cut journal funding.
(4) The government’s incentives are not aligned with ours. They don’t particularly want to advance mathematics for its own sake, but rather to the extent it helps achieve other goals. I don’t want the government micromanaging our publishing system according to their own priorities (for example, increasing funding to journals in areas they care more about or that sound good to voters).
Ultimately, when I see the EPSRC disasters or massive cuts to state universities in the US, I’m really glad these organizations are not in charge of our publishing system. The sort of decentralized system we have now is better than any centralized system I can envision.
Elsevier say that there is no viable economic model for the journal industry save the existing one.
I haven’t seen them say that, and they claim to be open to other funding models (see, for example, http://poynder.blogspot.com/2012/02/elseviers-alicia-wise-on-rwa-west-wing.html). I don’t expect them to lower their prices without a fight, but I doubt they are especially committed to their current business model, except to the extent that it seems to be working for them.
short-circuit the whole system by having money flow directly from funding agencies to publishers…
That’s part of what worries me: if all the money flows through one place, then that can be a single point of failure, so it’s very important to make sure it remains under control of the research community (rather than being manipulated through politics). Lobbying to a handful of government agencies is much easier than to thousands of universities.
You might be interested in SCOAP3, which is a consortium that will sponsor open access journals in high energy particle physics with no publication charges, starting in January, 2013. It sounds similar to what you imagine, except it is not a government-run project. It will get funding from a mixture of government funding agencies and university libraries, and it will redirect this funding to selected publishers for open access journals. Elsevier is one of those publishers.
Doing this in math would be a lot harder, since math is a vastly larger field than particle physics, but one could start small and build out from there. The role of funding agencies and large organizations (e.g., particle accelerators) is much lower in mathematics, so I expect libraries would play an even bigger role in a math analogue of SCOAP3. The hard part is getting it started, which has been a multi-year process for SCOAP3. I hope the IMU can play a leading role in organizing this, and they are well aware of the possibility, but it will take real thought to figure out what is feasible and desirable.
Is SCOAP3 directed toward existing journals made free or toward creating new journals. If it is creating new journals than why to distribute to so manz platforms and commercial publishers as well. New journals would be cheaper from a single large platform (like arXiv), or even overlay for the existing arXiv. It is waste of money to get lots of small expensive sites, instead of doing one big whose price is comparably as small (per unit size) as the funding of arXiv. The small changes are the biggest enemy of big changes.
SCOAP3 works with existing journals. The huge advantage it has over more radical proposals is that it’s more or less risk-free and a Pareto improvement (it offers real advantages without actually making anybody worse off). Even so, it took a lot of negotiation to set up: you have to convince publishers that they really aren’t taking a risk or setting a bad precedent, you have to convince libraries that it is better to fund this project than to wait and hope someone else does, you have to deal with technicalities (for example, state universities are sometimes not legally allowed to buy anything except for the lowest available price, out of concerns about corruption, so you need to offer participants a governing role that justifies paying money for a journal that will then be freely available to everyone), etc. However, they have managed to work it all out. If they were aiming for something riskier or more controversial, I doubt they could have done it at all.
The way I see it is that the perfect is the enemy of the good. SCOAP3 is certainly not the best of all possible publication systems, but it may be the best we can reasonably get in the foreseeable future. And the most important aspect from my perspective is that it’s the only scalable business model we currently have for full open access (with no time delay) and no publication charges, without substantially changing what we expect from a journal (in terms of copyediting, formatting, archiving, etc.) or relying heavily on volunteers (which can’t be scaled up a hundredfold). So I see sponsoring consortia like SCOAP3 as the idea that could save us from publication charges if the outside pressure towards open access gets too strong.
[I think open access is great, but math is already doing an excellent job at “green open access” through the arXiv. From my perspective, subscriptions with guaranteed open access after five years and freedom to post to the arXiv would be better than full open access based on publication charges, and a SCOAP3 analogue would be better than either one.]
Ok, I like the idea, thanks for explaining. My negative remarks were about a possible nonoptimal scenario for creating new journals on a variety of platforms.
1 to 21 of 21