Monday, January 28, 2013

Who Pays in Author Pay?



I have read several comments recently indignantly protesting that "author-pay-journals" are not after all actually "author-pay". The fees the journal charges for publication of an article are supposed to be called "article processing charges", and there are supposed to be safeguards in place to guarantee that articles are published whether or not the fees are collected. These are safeguards offered with the best of intentions. They are hopelessly naïve and perhaps even slightly dishonest.

The new open access mathematics journals offered by Cambridge University Press are a good example. Proponents claim that once a paper is accepted for publication, an author has merely to arrange for his/her institution to send a letter stating that it is unwilling to pay the article charges – with no need for justification whatsoever. And I am quite certain that in the heady early days of gold open access, some institutions and funding agencies will indeed pay the charges. But some will not, especially without required justification. Over time, will the former continue to subsidize the latter? Surely not, and surely a business model built on volunteer payments alone cannot sustain itself in the long-run.

Several years ago, earnest scholars who had begun free online journals ran out of volunteer time and the modest amount of money needed to keep their journals running. These were high-quality journals with admirable goals, and the scholars proposed a simple scheme to keep them going: They asked libraries to pay a very small subscription fee, hoping that many small fees add up to the modest support they needed. That small fee would be a contribution to the battle against high subscription fees for commercial journals, they argued. Surely libraries would want to contribute. Invariably, they did not. If universities could access journals for free, why should they use scarce dollars to pay for access?

Now we are expected to believe that universities will make a much more substantial voluntary contribution to publish the papers of their own faculty. Maybe, for a while, they will. But over time, as dollars become scarce and there is pressure to spend them on other important things, even wealthy universities will decline to subsidize less-wealthy universities.

The long-term consequences of these policies are determined partly by simple business principles, but even more so by basic human psychology. Alas, the naïve scholars who assure us that gold open access can be accomplished without authors ever actually paying the charges are not especially attuned to human psychology (or business principles). The publishers who are their partners in the new journals are not naïve, however, and the initial low article charges and assurances that no one will actually have to pay are slightly dishonest. They know better. This is simply bait-and-switch, albeit in a scholarly and refined setting.

The author-pay model is exactly that—the author pays. We should not try to obscure reality with fanciful promises. Right now, the fee may come from the university or some funding agency, but inevitably the author "authorizes" the charge. For many new and proliferating journals, authors themselves are indeed paying, and journals cannot prevent them from doing so. Over the long-run all journals will require payment somehow—they must. If one insists on gold open access, this is the price one pays. It may be worth the cost, but pretending there is no cost is foolish.
John Ewing

Saturday, November 10, 2007

What is the price of journal?

What’s the price of a journal?

For that matter, what is the price of a car or a novel or a loaf of bread? All these things are frequently discounted, but we don’t throw up our hands and claim that they don't have a "real” price. Yet on several occasions recently, I’ve heard people say that we can't tell the price of journals because they are often discounted.

When the editorial board of the journal Topology resigned and began a competing journal, Elsevier wrote: “Because the majority of our subscribers purchase this journal in a larger set of journals, most are paying a fraction of the institutional subscription price.” I’ve heard similar arguments from other publishers, who like to compute the “price” of a journal by dividing the total revenue by the number of “subscribers”. But that’s not the price! It’s the “average revenue per subscriber."

The (list) price of a journal is set by the publisher, and it’s plainly visible to anyone who examines annual price lists. For some journals, there may be a two-tiered price, one for institutions and one for individuals, but in every case there is a price. Just as for cars or novels or bread, journals may be sold at a discount. But it's important to remember that publishers discount journals for business reasons, not because, in a sudden fit of remorse, they want to lower the price. Journals are sometimes discounted to agents, who consolidate them to help libraries purchase from multiple publishers. They are discounted to institutional members of scholarly societies as a member benefit, in return for dues. And journals are discounted to subscribers who buy bundles of journals, often making a commitment to buy for several years. In each case, the publisher is discounting journals in order to gain some advantage -- it's a simple business arrangement.

There is nothing wrong with discounting journals; it's good business. But it doesn’t change the price. Indeed, the price is the starting point for all discounting arrangements, defining the terms of a bargain: I’ll return a portion of the price in return for some action on your part – consolidating, being a member, or purchasing a bundle. Confusing the discounted price with the actual price ignores one half of the bargain.

We should pay attention to the list price of a journal because inevitably some subscribers (quite often, most) pay the list price. But there are other reasons not to let publishers substitute the "average revenue per subscriber" for the price. The average revenue is a quotient, and publishers control both the numerator and the denominator.

Unlike the list price, we must rely on the publisher to tell us the numerator, that is, the total revenue for a journal. Calculating total revenue sounds straightforward until one realizes that when selling bundles, large publishers apportion revenue among many journals – a somewhat mysterious process that isn’t easily discovered. For many publishers, the total revenue assigned to a particular journal is a very fuzzy number indeed.

The denominator is even more problematic. How many subscribers does a journal have? If a publisher adds many journals to bundles at no charge, the number of subscribers will quickly rise. But adding unwanted (and frequently unused) journals to bundles doesn't really change the number of subscribers to each journal in any meaningful way. Allowing publishers to use these arrangements to calculate either the average price per journal (for an institution) or the average revenue per subscriber (for the publisher) is like allowing politicians to count all those people who might have voted for them (but didn't vote) in an election. And allowing publishers to tell us the "real" price of their journals is like asking car salesmen to dictate the "best" price for their cars.

Scholars face a crisis today caused by high journal prices. If they are going to make headway in addressing that crisis, they have to get smarter about journals and more sophisticated about business practices. They can't allow publishers to redefine the problem by redefining the price. That's neither smart nor sophisticated.

John Ewing

Thursday, November 1, 2007

Where are Journals Headed?

Where are Journals Headed?

Why we should worry about author-pay

People often ask me what I think will happen to journals. I’m a mathematician, but I know I won’t find the answer there. Mathematics journals account for roughly 5% of scholarly journals[1]-- journals in the biological and medical sciences account for about 50%! And this has consequences: At nearly every meeting about journals, biomedical journals dominate the discussion.

The business model adopted by biomedical journals will most likely be the model adopted by all others because biomedical journals not only dominate numbers but revenues as well. Customers (in this case, institutional libraries) don’t like to deal with multiple business models. They don't easily divide budgets into new pieces (say, subscriptions versus page charges), and they don’t like making complicated purchasing decisions. Dominant products and services usually shape those of lesser importance, and biomedical journals are clearly dominant.

The real question is therefore where are biomedical journals headed, and the answer seems obvious: They are moving towards an open access, author-pay model -- one in which journal content is available at no charge to everyone, but authors pay a fee prior to publication. To many biomedical scientists, this feels like the right model. The purpose of publishing a biomedical paper is to make the results available now, not to preserve them for the future. The focus is on immediacy. Paying a "posting-fee" makes sense, and since most biomedical research is supported by grants (often large ones), a relatively small posting-fee is easily absorbed in the grant. This is the model underlying experiments such as the Public Library of Science, and it is the model implicitly promoted by the National Institutes of Health in the U.S., as well as by various biomedical organizations throughout the world. It is the logical successor to the subscription model when papers are made freely available after only a six month embargo.

Should we worry that all scholarly journals may follow a course dictated by one discipline's need for immediacy and availability of ample grant funds? Some open access proponents claim not. Everyone wins, they say, because not only do we gain universal access but, if the posting-fee is only the cost-of-posting, we will also save money -- lots of it. As for the lack of grants, institutional budgets will merely shift from subscriptions to “page-charges” (that is, author fees), so that even those without grant funds will be able to publish their research. It's simple, they say, a model that benefits all.

But there are good reasons to worry about this sanguine view of the new model for journals.

(i) In areas where most research is not grant supported, universities and colleges will have to pay author-fees by reallocating money from libraries (subscriptions) to other parts of the institution (departments? divisions?) that need the funds. But reallocating money is never a simple process. Will those who pay author fees from grants (biomedical sciences) be willing to give their library budgets to those who cannot (say, the humanities)? I suspect not. Will administrators look for ways to save money by shifting funds to other uses? Long experience suggests they will. Will departments with prestigious faculties demand more of the funds than those with less prestigious? Of course they will, and this will exaggerate differences throughout the university. Various constituencies will vie for funds, with inevitable winners and losers. Perhaps that's not bad, but it's surely not "simple".

(ii) The change in who makes decisions will change the market; this is basic economics. In the subscription model, users and librarians make decisions; in the author-pay model, authors and publishers make them. To succeed in the subscription model, a journal must secure enough subscriptions by convincing users and librarians that it has intellectual value. To succeed in the author-pay model, a journal must convince enough authors to submit papers and then it must accept enough of them to make money. Price will vie with prestige. The most prestigious journals will charge more and will attract authors who can pay the cost (grants will help). The less prestigious journals will discount their price in order to attract more authors and will increase the acceptance rate. Some institutions may demand that scholars use less expensive journals; others will demand that their faculty publish only in expensive ones. The result will be a distorted and ugly market, driven by some of the same forces that drive vanity publishing. This is what happens when a market is driven by producers instead of consumers.

(iii) The author-pay model emphasizes immediacy. All money exchanges hands before the article appears when the author pays a "posting-fee". After a short period of time, the material in the journal has no monetary value to the publisher, other than to attract more authors. This is a subtle but profound change from the subscription model. Because anyone can post articles on the web, unscrupulous publishers will take advantage of this short-term view by accepting marginal papers (or just plain junk) into newly created journals in order to make easy cash. Those who think scholars will not publish in such "instant journals" have not looked at current marginal publishers (who are kept in check only because they have to convince someone to buy their publications). Almost surely, more papers will be published in such a system.

(iv) The large commercial publishers will thrive in this new model. In fact, all large commercial publishers already have units devoted to open access publishing and are (quietly) pushing the author-pay model. Why? They will now produce a product for which they get paid by the supplier, in advance, without risk, and with lower overhead (because they don't have to sell subscriptions). And because the large publishers are diversified, they can take advantage of a changing environment. Small journal publishers in areas that have no grants to pay author fees will quickly go under; large publishers will expand into areas that are most lucrative. Large commercial publishers will end up with less competition in a market that is more easily manipulated -- a market they will dominate even more than now. Of course they are pushing the author-pay model!

The fundamental problem for journals is simple – we pay too much for them! It’s not access (which has never been better). It’s not our business model (which is shared responsibility). It’s not how we pay but rather how much we pay!

Many proponents of the author-pay model think we can solve this problem by switching to a new business model. Some have faith that publishers can be persuaded to set author fees only slightly higher than publication costs. But publishers who have profited from subscriptions in the past will certainly expect to profit from author fees in the future. (I can assure you that commercial publishers have this expectation.) Others believe they can run inexpensive author-pay journals themselves to compete with established journals, miraculously succeeding with upstart author-pay journals where upstart subscription-based journals have failed in the past. But there is no basis for this optimism. Indeed, since we will likely publish far more than ever before, we will likely spend far more as well. And here’s the largest worry about the author-pay model: It does not solve the fundamental problem of journals -- this model makes it worse!

We are therefore heading in the wrong direction. Scholarly journals are sick and they need attention. But instead of following a regimen of reasoned and disciplined remedies -- instead of driving down prices by the steady, concerted actions of authors, editors, and librarians -- we are bleeding the patient with open access models, trusting in miracles (that university administrators will shift funds from those with research funds to those without), and praying that publishers will repent their ways.

It is ironic that those leading us down this path of folk remedies and faith healing come from the biomedical sciences.

John Ewing



[1] These percentages are derived from data in the Genamics JournalSeek database. See http://journalseek.net/information.htm .