Cites & Insights: Crawford at Large
ISSN 1534-0937
Libraries · Policy · Technology · Media

Selection from Cites & Insights 4, Number 4: March 2004

Library Access to Scholarship

Walt Crawford

You could consider this the new running head for Scholarly Article Access. That may change, but at least the new name emphasizes the aspect that concerns me most: The ability of libraries to maintain short-term and long-term access to (and collections of) the scholarly record.

The Tipping Point Continued?

I’m probably being too optimistic in viewing recent events as a form of tipping point toward the decline of the big deal and the stranglehold of big international STM publishers on academic library budgets. But I’m an optimist by nature. The May 2005 “Crawford Files” in American Libraries offers my own brief description of the future I’d like to see. Meanwhile, a few notable events since the January summary, in chronological order.

University of California

On January 7, the eleven university librarians of the University of California (including the 10th campus now being built and the California Digital Library) sent a letter to the UC Faculty Senate reporting the successful conclusion of negotiations with Reed Elsevier. Briefly, a new five-year contract provides access to 1,200 of Elsevier’s titles from all divisions, dropping 200 titles (formerly included in the big deal) that weren’t selected by any campus. The letter notes that the new contract has “arrested for now the price inflation that has been common in this market” but does not state the price. It goes on to discuss the ongoing need to address issues in scholarly communication, including this paragraph:

The economics of scholarly journal publishing are incontrovertibly unsustainable. Taming price inflation is not enough. Unless we change the current model, academic libraries and universities will be unable to continue providing faculty, students, and staff with the access they require to the world’s scholarship and knowledge. Scholars will be unable to make the results of their research widely available.

Ongoing action includes UC library work to stretch collections dollars through consortial licensing, inform librarians and faculty on the dimensions of the scholarly communications crisis and possible ways to address it, and support “alternative means for publishing scholarly materials that make high-quality peer-reviewed work available at an affordable price.”

Clifton B. Parker at UC Davis provided additional information in a January 6, 2004 report. The contract includes systemwide online access and a single print copy of each title to be managed in a regional facility, so that campus libraries can cancel Elsevier print subscriptions without depriving faculty of access to print journals as needed. “The net result is that the UC libraries will spend 25 percent less on Elsevier subscriptions in 2004 than they did in 2003 ($7.7 million in 2004 as opposed to $10.3 million in 2003).”

Triangle Research Libraries Network

A story posted January 14, 2004 on Library Journal’s website notes TRLN has announced it will not renew its Elsevier deal and calls this “another blow to the big deal.” TRLN officials said the decision not to renew the deal followed months of unsuccessful negotiation. I mentioned North Carolina State University’s faculty resolution opposing big deals last time around; Duke and the University of North Carolina at Chapel Hill form the rest of TRLN.

TRLN and member libraries had two principal objectives in the negotiations:

1. To regain and maintain control over library collecting decisions in order to meet the constantly evolving information needs of faculty, researchers, and students; and

2. To manage overall costs in order to keep Elsevier expenditures consistent with materials budgets that have not been increasing at anywhere near Elsevier’s annual inflation rate.

Elsevier’s final offer fails to meet both of these objectives.

Specifically, Elsevier insisted that libraries commit to zero cancellations, making collection policy inflexibly and “inordinately privileg[ing] the journals of a single publisher.” The memo notes that this would create a growing imbalance in collections and be detrimental to scholarly associations and society publishers. The problem was magnified by Elsevier’s insistence on “significant annual cost increases above TRLN’s current contract terms,” which amounted to more than $4.5 million per year.

The memo anticipates the loss of access to 400-500 journals per campus and the need for campuses to cancel more local subscriptions. It goes on to note that the Elsevier situation is “only the most extreme symptom of a much larger problem.” As with UC (which the memo cites along with Harvard and Cornell), this is only the beginning:

We firmly believe that universities must respond to this economic crisis of the state of scholarly communication. Libraries must be empowered, through dialogue with the university community, to obtain appropriate research material without sacrificing content and budgetary decisions to the publisher. Future library negotiations should follow the principles adhered to in this particular process, that libraries must make collection decisions and manage costs.

After some details, here’s the telling conclusion: “[The libraries] will begin to explore with you new models of scholarly communication that may, in the long term, help reduce costs and make scholarly information more widely available.”


In an MIT Libraries memo downloaded February 6, the libraries note steps taken to “reduce the impact of two large commercial publishers on our ability to make responsible decisions in selecting information resources for use at MIT. Specifically, we declined three-year renewal contracts that would have required us to guarantee ongoing spending levels with Wiley InterScience and Elsevier Science.” MIT was offered a three-year renewal of big deals through NERL, the NorthEast Research Libraries Consortium. The two packages combined constitute one-third of MIT Libraries’ budget for serials (print and online alike) and include the usual zero-cancellation policy (or an alternative requirement to buy an equivalent new serial for each cancellation).

MIT took one-year deals, taking a hit in the process. Elsevier charges more for a one-year package. The Wiley package costs about the same but won’t include as much content. Both decisions provide MIT with flexibility to make decisions in the next two years “based solely on the specific needs of the MIT user community, without giving unfair advantage to specific publishers.”


A February 5, 2004 article in the Harvard Gazette, entitled “Libraries take a stand,” notes that students and faculty logging on after winter break found fewer Elsevier periodicals. Director Sidney Verba’s comment on the decision to eliminate some Elsevier publications: “It was driven not only by current financial realities, but also—and perhaps more importantly—by the need to reassert control over our collections and to encourage new models for research publication at Harvard.” The article notes similar actions at Cornell, TRLN, and Johns Hopkins.

Some of the cuts were duplicate print subscriptions, as Harvard works more effectively to minimize needless duplication within its extensive library system. As with MIT, Harvard also arranges some consortial licenses through NERL—and decided not to take the NERL Elsevier license this year. The article goes on to quote Markus Meister, who serves on the PLoS board and discusses the need to change the structure of scholarly communications.


On February 9, the University of Connecticut Faculty Senate passed a resolution concerning access to the scholarly literature. “The business practices of some journal publishers [are] inimical to [access to the scholarly literature] and threatens to limit the promise of increased access inherent in digital technologies.” Noting that the rising cost of journals and databases increasingly constrains library collection development, the resolution calls on faculty, staff, students, and administrators to “take greater responsibility for the scholarly communication system.”

The resolution “encourages senior tenured faculty to reduce their support of journals or publishers whose practices are inconsistent with the health of scholarly communication” through the usual means: submitting fewer papers, refereeing fewer papers, resigning from editorial posts. It’s a gradual call, not a plea for complete boycott. The resolution calls for university groups to “reward efforts by faculty, staff, and students to start or support more sustainable models for scholarly communication” and adds supporting language.

Donald Knuth

This is a slightly different situation—the latest in a small but growing number of cases, perhaps higher profile among those (like me) who treasure Professor Knuth’s work in a range of fields

Knuth sent a 14-page single-spaced letter to the editorial board of the Journal of Algorithms on October 27, 2003. He is a founding editor of the journal and has been involved with it throughout its nearly 50 volumes.

When founded, the journal was published by Academic Press, with whom Knuth has been involved to some extent since 1965. When Harcourt Brace Jovanovich acquired Academic Press, Knuth asked them to “do their best to minimize the [pricing] effect on libraries,” and says they did so during the next few years. Toward the end of the 1990s, however, the price started to increase fairly rapidly—and when Reed Elsevier purchased Harcourt, the increases continued or accelerated. (Knuth gives a chart showing not only the annual subscription price but also the price per page. It’s worth noting that even the 2003 price for the Journal of Algorithms, at $700, is low enough to be under the radar for many academic libraries—but that’s almost double the 1997 price.)

There’s a lot more to the letter—as Knuth admits, he’s never learned how to be brief, a failing with which I can identify. He believes the price per page should have dropped, since few mathematical journals still require typesetting or keyboarding—but it’s doubled, not only in this journal but in “virtually every other mathematics journal produced by commercial publishers.” He notes that journals produced by nonprofit organizations have generally kept costs steady. Notably, Knuth makes a direct connection between high prices and libraries: “My blood boils when I see a library being overcharged.” When he wrote a letter to Elsevier, it did what you’d expect: Ignored the letter and did not reply.

This is a very long letter. It includes quite a bit about the STM crisis, Open Access, alternative models, what have you; I wouldn’t even attempt to summarize the whole thing. It’s worth noting that Knuth likes print, in part because it’s the easiest way to browse through a range of articles. The letter is well worth reading and is available at

Knuth wanted the editorial board to vote on possible futures for the journal—to stick with it as is, switch to a nonprofit publisher (e.g., a university press), affiliate with a learned society (probably ACM or SIAM), or move to university hosting as a true open access journal.

The editorial board resigned. The new ACM Transactions on Algorithms will be launched with the same editorial board. Elsevier is establishing a new editorial board and plans to continue publishing Journal of Algorithms.

Which, unfortunately, only goes to illustrate a key problem with alternative models: They do nothing to ease pressures on library budgets unless they either replace existing journals, cause those journals to become irrelevant, or weaken those journals enough that the publishers lower prices.

BNP Paribas

This equities firm issued a lengthy report on “the impact of the development of new communications technologies on the global professional publishing industry.” I’m not sure where or whether you can find the report, and it’s very much an investment report—but it’s also fascinating reading. The firm estimates professional publishing as a $40 billion industry worldwide, $20 billion of that in the US.

This report claims that first-copy costs (everything not directly associated with print publishing) represent 85% of the total costs of STM journals. I’ve claimed for some time that costs directly associated with print publishing represent only about 15% of the price for most books, but I’m a little surprised to see such a claim for cost (a very different animal than price) and for journals. The same paragraph also notes that median circulation per journal has fallen from 2,500 in the late 1900s to 1,900 at present, and that libraries now account for 85% of the sales of academic journals in the U.S.

I don’t understand this claim: “The serials crisis peaked in the late 1990s.” While publisher price inflation may have peaked then, I’d be hard put to find librarians who believe that the pressures on their budgets have eased in recent years!

The report has reasonably good commentary on open access publishing and some of the pressures around STM journals. It notes that some OA journals have already achieved very high impact factors. As you’d expect from an investment house, it dismisses the current publication charges for BioMed Central and PLoS as unsustainable, suggesting that $2,000 to $2,500 is a more plausible level—and even that would be an enormous improvement over Reed Elsevier’s current $4,400 revenue per article.

One chart is particularly telling: The estimated operating margins (“gross profit”) for various segments of publishing. STM journals run 35 to 40%; consumer magazines and book publishing, on the other hand, are both in the 8 to 11% area, with trade magazines only a bit higher.

Cries and Alarums

Established publishers won’t let OA grow without casting every possible aspersion. imi insights for October 2003 includes Kate Worlock’s “Open access: A step back in time,” based on an interview with Elsevier’s Arie Jongejan. Jongejan is eager to challenge “myths and misconceptions surrounding the emotive area of open access.” What are those myths and misconceptions?

First, access: Jongejan claims, “Around 70% of the audience which might be interested in accessing Elsevier’s scientific, technical and medical content can at present do so.” So, as long as libraries cough up those ever-higher fees for ScienceDirect, there’s no access problem. Clear enough?

Second, “the perception that open access is a free and egalitarian business model.” Jongejan claims OA discriminates against authors based on their ability to raise funds.

Third, “the underlying assumption that the current publishing process adds very little to the content being published.” Jongejan mentions refereeing and peer review as important added value—and says he “believes that this is not always the case with some open access players, who take the role of review much less seriously.” He quotes a BioMed Central referee policy to back his assertion—although it’s very hard to read the quote as being less than proper refereeing.

“Jongejan does not believe that open access is either economically sustainable or more efficient than traditional publishing models” and claims OA publishers “will need to demand between $3000 and $4500 per article to cover publication costs.” Note that Elsevier currently averages $4,400 revenue per article, including 40% pure profit.

 “Submitting with a cheque potentially compromises the review process,” and OA journals “can therefore be seen more as an author exposure service than a publishing operation, with a potential lack of peer review jeopardising quality control.”

Jan Velterop of BioMed Central had some pointed comments about this piece, beginning: “It is never a good idea to throw stones if you live in a glass house. Especially not if you don’t understand your target.” He says Jongejan’s statements “stretch Jongejan’s—and Elsevier’s—credibility to [the] snapping point.”

Do you need Velterop’s comment about Jongejan’s absurd assertion that access isn’t really an issue? “First of all, 70% is not the 100% it should be, and secondly, how would he define ‘the audience which might be interested…’? It betrays supreme arrogance to pretend that one knows, as publisher, who might be interested in the research articles that are published…” Velterop leaves out one key fact: The 70% is based on an unsustainable Big Deal system. Libraries simply cannot afford to provide access under those terms.

Velterop pushes data mining as an advantage of OA. I don’t know enough to comment, so I won’t. He does point out that the second myth is mostly nonsense—it’s almost always institutions rather than authors that pay, and OA publishers frequently waive fees for authors from developing countries.

Third is the disingenuous attack on the editorial integrity of OA journals. Velterop notes that the phrase used by Jongejan to suggest BioMed Central doesn’t require originality is actually intended to do the opposite—to prevent “salami slicing” (least publishable unit) articles.

And, of course, Velterop points out that Jongejan’s claim on actual OA costs is nonsensical. “Given their profit margins, Elsevier’s own cost per article must be well below $3000. And that includes print, postage, discounts to subscription agents, elaborate access-control measures, subscriber databases, sales forces…at least 65 offices, massive inefficiencies, plus, according to Jongejan, a cost of between $0.05 and $0.15 per download.”

The January 30 Chronicle of Higher Education includes “The promise and peril of ‘Open Access’” by Lila Guterman, with two sidebars. She discusses PLoS and its promise—but also some of the doubters. I have a lot of trouble with the first doubt (the claim by scientific societies that it’s appropriate for their activities to be underwritten indirectly by libraries), but agree with the second, at least as a potential issue:

What’s more, open access may not even save universities money. If the new publications multiply but do not immediately replace subscription-based journals, the transition period will be uneasy and expensive—and no one knows how long it will last.

“If we have to pay for both the existing journals and the author-pays fees, we’re going to get killed,” says Charles E. Phelps, provost of the University of Rochester.

Then she discusses a few budget realities—like the Journal of Comparative Reality at $18,000 a year, Brain Research at $21,000, or Nuclear Physics A and B at more than $23,000. She quotes Carol Tenopir: “The subscription model where the library pays is beginning to break down.” And she goes on to mention several of the cases where faculty are now backing libraries in reducing subscriptions—Duke along with Cornell and UC. Naturally, Reed Elsevier is “trying to be as sympathetic as we can be.” And the provost at Carnegie Mellon says library woes are still not “very front and center” for faculty members at many universities: There’s a lot of education to be done.

The OA discussion is long and interesting, including some figures for Duke that are a bit startling—e.g., that PLoS-level author fees for scientists and social scientists could cost the university more than its entire current budget for serials, including those in the humanities and medical center. Naturally, Arie Jongejan is heard from. Bizarrely, AAAS says it would have to charge $10,000 per paper for Science to become OA because it has such a high rejection rate—but does anyone really expect Science or Nature to convert to the OA fee model (as opposed to providing open access to published articles)?

The article ends on a sad note that indicates just how much education is needed. Ricardo Pietrobon at Duke prepared a manuscript a couple of months ago that he intended to submit to a BioMed Central journal. “But, warning him that the journal might fold, colleagues at Duke talked him out of it. He sent the paper to a traditional journal instead.” Sigh.

One sidebar is truly strange, in commenting on the fear of traditional publishers that governments may legislate change. Here’s a statement from the International Association of Scientific, Technical, and Medical Publishers: “Abandoning the diversity of proven publishing models in favor of a single, untested model could have disastrous consequences for the scientific research community.” To which a natural response might be, “What diversity is that?” At least for the “moderates,” those who believe that OA journals can and should replace many, perhaps most, traditional STM journals, but neither will nor probably should replace all of them, OA adds diversity—particularly since there are many different kinds of OA journal.

Highlights from Peter Suber

If you care about OA, you should subscribe to the SPARC Open Access Newsletter. If you do, you’ve seen the items I’ll mention here. If you can locate Suber’s January 29 colloquy on “The promise of open access publishing,” held by the Chronicle of Higher Education, you should. Suber handled friendly and hostile questions as adroitly as you’d expect.

In Issue 69 (January 2, 2004), Suber notes some of the highlights for OA in 2003, including the shift of objections to OA “from ideology to science”—that is, that most questions and objections are now ones amenable to empirical investigation. That’s as it should be. He also offers interesting perspectives on the virtues and drawbacks of proliferating copies. I’m not all that interested in the downside that “copies interfere with the measurement of traffic and usage,” particularly since I suspect that some of the very high OA usage numbers are meaningless anyway. As he notes, the only measures of impact that really work in a paper environment—citation analysis—aren’t affected by multiple copies.

Issue 70 (February 2, 2004) includes Suber’s “predictions for 2004,” an ambitious set of 14 predictions. There are one or two I might quarrel with (now there’s a surprise), but not many more than that. I’m sure he’s right that there will be “more struggle over the exact definition of the term ‘open access,’” particularly given some of the personalities involved—and I’m also sure he’s right about this one: “There will be less unity in the OA movement, or at least less concern to preserve solidarity in every public discussion.” The question, then, is whether that’s a bad thing. I would argue that it’s a very good thing, and that it would be even better if OA moved from being a “movement” to being several growing and diverse subsets of an increasingly diverse publishing landscape. But then, I’m not that fond of movements (and am not ready to lump OA, open source software, copyright, and censorware into an “information commons movement,” as Suber apparently is).

Issue 70 also includes an interesting essay on open access in the humanities, a fascinating essay including nine significant reasons that OA won’t grow as rapidly there as in STM. I’m keeping the list for further consideration; to summarize it here would take almost as much space as Suber’s clear, tightly written original.

Cites & Insights: Crawford at Large, Volume 4, Number 4, Whole Issue 47, ISSN 1534-0937, is written and produced at least monthly by Walt Crawford, a senior analyst at RLG. Opinions herein do not reflect those of RLG. Comments should be sent to Cites & Insights: Crawford at Large is copyright © 2004 by Walt Crawford: Some rights reserved.

All original material in this work is licensed under the Creative Commons Attribution-NonCommercial License. To view a copy of this license, visit or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.