Contact Us | Members Only | Site Map

Association of Research Libraries (ARL®)

  Resources Contact:
Lee Anne George
Publications, Reports, Presentations
ARL: A Bimonthly Report

A Snapshot in Time: ARL Libraries and Electronic Journal Resources

Share Share   Print

by Mary M. Case, University Librarian, University of Illinois at Chicago

For the past decade, ARL has been tracking its members' expenditures on electronic resources. Between 1994/95 and 2001/02, expenditures on electronic resources for the typical university research library have grown almost 400% to almost $1.4M (see Figure 1), while the overall library materials expenditures have grown only 61% (see Figure 2).(1) Electronic journals account for the greatest proportion of the electronic expenditures claiming 92% of these dollars in 2001/02. Spending on electronic journals has increased 712% since 1994/95 (see Figure 3). Electronic journals now account for 26% of a library's overall serials expenditures, up from 5% in 1994/95 (see Figure 4). Expenditures on electronic books or other one-time purchases account for less than 4% of current monographic expenditures.

Given the significant and increasing investment research libraries are making in electronic journals (over $154M total reported by 108 libraries in 2001/02), ARL undertook an extensive survey of its members' electronic journal subscriptions in fall 2002 to better understand the issues libraries are facing in ensuring that electronic resources can be used effectively on campus. In addition, another survey was conducted in 2003 at the request of members to ascertain how libraries were dealing with their journal package renewals and what licensing issues were of greatest concern to the community.

The Surveys

ARL collects extensive amounts of data from its members every year. This information is essential in identifying general trends in library collections, expenditures, and services. For some circumstances, however, these statistics are not detailed enough to provide answers to specific questions. For example, in 1997, Reed Elsevier and Wolters Kluwer (the two largest science, technology, and medical (STM) publishers in the world at the time) announced their intention to merge. The Department of Justice (DoJ) contacted ARL to assist in the investigation of this merger, but the Association did not have the detailed level of data to help with its analysis. In spring 1998, however, more than 50 ARL libraries combed through their records (often manual) and provided specific data on more than 3,000 journal titles. This information helped the economist at the DoJ develop and test a theory that demonstrated the dangerous market power of STM publishers and the significant price increases that result from mergers in this industry.(2) In 1998, however, the data was all based on the world of print journals--that world was about to change.

While Reed Elsevier and Wolters Kluwer called off their proposed deal, many publishers have merged since then. Reed Elsevier purchased Harcourt General in 2001, and Candover and Cinven purchased BertelsmannSpringer in 2003 with the intention of merging it with Kluwer Academic, which they purchased earlier the same year. In both of these instances, the companies argued that the electronic environment was changing the dynamics of the market with clear potential for greater competition from new entrants and more aggressive action to create new systems from academics themselves.(3) They argued that competition authorities ought to let this new world evolve before attempting to block a merger or regulate the industry. The competition authorities apparently agreed and allowed these two major transactions to proceed. It was clear that ARL needed to start gathering data on electronic journals to better understand, what we perceived to be, the increasingly detrimental effect of mergers in the electronic environment.

To that end, ARL developed a survey in 2002 to gather information about members' subscriptions, expenditures, and licensing terms for 60 publishers. Given the enormity of this task and to encourage participation, ARL identified a core of 12 major commercial companies and a subsequent set of 16 commercial and not-for-profit publishers on which participants could focus their data collection. Forty libraries responded to at least some portion of the survey. This paper focuses on the seven commercial and seven not-for-profit publishers whose electronic titles were most subscribed to by the responding libraries (see Table 1 for a list of publishers).(4)

The survey sent out in 2003 was intended to gather information on additional issues in licensing, particularly how libraries were thinking about their upcoming negotiations for many of their electronic journal packages. Fifty-seven libraries responded to this survey, and the results were previously shared with ARL members.

The following analysis proceeds by topic and includes responses from both surveys, as appropriate.

Subscriptions, Costs, and Pricing Models

One of the prime purposes of the 2002 survey was to get a sense of how many titles libraries were subscribing to from each publisher, what they were spending, and under what pricing model. Table 1 displays the average number of titles subscribed to in both print and electronic form from each of the 14 publishers. (Note the presence of Academic Press, which had only recently been purchased by Reed Elsevier, and not yet folded into Elsevier Science. In addition, Kluwer Academic and Springer had not yet merged.) The first column indicates the number of libraries that reported whether or not they subscribed to electronic journals from these publishers. Not surprisingly, libraries subscribe to many more titles from commercial publishers because they actually publish significantly more titles than the not-for-profits. In general, libraries are subscribing to more electronic titles than print. What one cannot tell from these numbers is the proportion of available titles that libraries are subscribing to from each publisher. Given the significantly lower production of the not-for-profit publishers, one could assume that libraries are subscribing to a greater proportion of their titles than the commercial publishers' titles. At this point, only a few libraries noted that they had canceled all of their print titles for a particular publisher. This topic is addressed in more detail below.

Table 2 shows the average expenditures on the commercial titles. The average print cost of the commercial titles ranges from $414 for Blackwell to $2,572 for Elsevier. In terms of electronic titles, the average cost ranges from $7 for a Taylor and Francis title to $742 for an Academic Press title. The extraordinarily low cost of the Taylor and Francis titles reflect its dominant electronic-free-with-print pricing model. For Academic Press, the higher cost of its electronic titles reflects its "flip" to an electronically based pricing model with print available for a surcharge. Given the variety of pricing models among publishers and the variety of pricing offered by any one publisher, it seemed useful to calculate an overall unit-cost-per-title-per-publisher that includes both print and electronic formats. For the commercial publishers, this unit cost ranges from $160 for Blackwell titles to $1,002 for Elsevier titles. On average, research libraries spent over $1.3 M on Elsevier journals in 2002--that's over five-and-a-half times the next largest expenditure of $240,831 on Wiley. Of the seven publishers included here, Taylor and Francis accounted for the least amount of expenditures at $85,730.

Table 3 shows the average expenditures on the seven not-for-profit titles. Of these, ARL libraries spent the most in 2002 with ACS, almost $42,000--that's less than half the amount with Taylor and Francis. They spent the least with ACM, only $5,772. The average print cost ranges from $64 for ACM titles to $1,771 for titles from the Institute of Physics. For electronic titles, the average cost ranges from $57 for a Project Muse title to $1,048 for an AIP title. The overall average unit costs range from $65 for Project Muse to $823 for AIP.

Table 4 summarizes the dominant pricing models of the publishers. Eight of the publishers offered options that included print, plus a fee for access to electronic equivalents; print, plus a fee for access to the full set of electronic titles; or electronic, plus a fee for print equivalents. Many publishers were still offering electronic access free with print, but this option seems to be disappearing in favor of options that secure greater rights for libraries, such as more content, archiving, interlibrary loans (ILL), e-reserves, and course packs.

Subscription Terms

1999 seemed to be a watershed year in the uptake of subscriptions to electronic journals (see Figure 5). By 1999, 50% or more of the respondents for nine of the publishers indicated that they had subscribed to electronic titles from those publishers. In fact, all seven of the not-for-profit publishers fell into this category, along with Academic Press and Springer Verlag. Libraries seemed most supportive of Project Muse with 50% of those subscribing reporting they had started their subscriptions in 1997. By 1999, 80% were subscribing. Blackwell, Elsevier, and Taylor and Francis broke the 50% mark in 2000, with Kluwer and Wiley following in 2001. This phenomenon is also reflected in Figure 2, which shows significant increases in spending on electronic resources in 1999 (30%) and 2000 (45%).

The most common duration of license agreements was one year. This included licenses with Blackwell, Springer, Taylor and Francis, AIP, Oxford, ACS, IOP, Project Muse, Cambridge, and ACM. Three years was the average for Elsevier, Kluwer, and Wiley, with everything from one to five years being reported for Elsevier. Academic Press was fairly evenly divided between one- and three-year contracts.

Five of the 14 publishers included clauses in their licenses that restricted the amount of cancellation of print a library was allowed during the term of the agreement. In essence, these clauses restricted the amount of money by which the total expenditure could be reduced. Seventeen libraries reported such clauses in the Wiley license, which allowed cancellations ranging from 0 to 10%. Seventeen libraries also reported cancellation limits in Blackwell's agreements, which ranged from 0 to 5%. (Although not statistically tested, there seems to be a positive correlation between a cancellation clause in a Blackwell license and the licensing of the resource through a consortium. This may be due to the granting of additional rights to a consortium, such as perpetual access in the Blackwell case that the publisher believes warrants the limitation on cancellations.) In the case of Elsevier, 11 libraries reported cancellation limitation clauses. These varied from 1 to 10% over the life of the agreement. Ten libraries reported no cancellations allowed by ACS, and nine libraries reported limitations by Kluwer of 0 to 6%.

As is well known, consortia are an essential partner for libraries in licensing access to electronic resources (see Table 5). For 10 of the publishers, a third or more of the libraries reported subscribing through a consortium. Kluwer attracted the most activity with 72% of subscribers accessing titles through a group arrangement, which was well ahead of the next nearest publisher, Academic Press, with 55% of libraries accessing titles through consortia. One respondent offers a possible rationale for Kluwer's success with groups: "...a small electronic surcharge and agreement not to cancel gives consortial access to many titles." OUP, AIP, Cambridge, and Taylor and Francis have very few consortial agreements--not surprising since their dominant pricing model is electronic access free with print.

One respondent to the summer 2003 survey characterized the importance of consortial relationships from the perspective of a library not in a consortium: "We have been hampered (we think) by the lack of an obvious consortial niche for our institution. We're exploring the possibility of joining one or more consortia for which we'd be a good fit and vice versa. We're very aware that consortia themselves are having a very difficult time with these issues and that joining one or more definitely will not be a panacea in funding e-journal collections. On the other hand, going it alone has been very tough."

In terms of methods of delivery, the content for 11 of the 14 publishers is accessed directly from the publisher's Web site. (There are exceptions for those libraries that access content through OhioLINK, which actually loads and serves electronic journals from its own site.) At the time of this survey, Blackwell actually had two separate companies publishing journals: Blackwell Science, whose titles were available on its Web site through a service called Synergy, and Blackwell Publishing, whose titles were available through third-party services such as Ingenta, EBSCO, and OCLC. Blackwell was merging the two companies, and respondents anticipated that the two separate licenses would be consolidated and all titles would eventually migrate to Synergy. Libraries also reported that some Oxford titles were accessed through the publisher's Web site and others through third-party services, while access to Taylor and Francis titles was primarily through third-party services.

Journal Packages

Multiple-year subscriptions, cancellation limitation clauses, and caps on annual inflation are the hallmarks of what has become known in our community as the "Big Deal." In the 2002 survey, 65% (or 20 of 31 respondents) of the libraries that indicated they were subscribing to electronic journals from Elsevier, indicated they also subscribed to ScienceDirect in a multi-year deal. Another 13% (4 libraries) noted they subscribed to electronic-only packages. Seven libraries (or 23%) indicated they had access to some Elsevier titles through Web Editions (a free electronic access with print program).

In 2003, 57 ARL members responded to a survey asking them about their plans to renew their package subscriptions. Almost 40% (22) of the libraries indicated they were planning to cancel or considering canceling a bundled package for the 2004 renewal. Respondents identified seven different packages in their list of products under consideration, including packages from not-for-profit publishers. While 73% (16) of the libraries evaluating packages cited budget difficulties as their primary motivation, other reasons listed included low use, high cost, and the desire to maintain flexibility.

Most of the libraries undergoing review of their packages employed processes that included communication with faculty, whether through formal targeted events or established departmental liaison contacts. Most libraries provided data to faculty on their budgets, product costs, and use.

While we have not followed up with ARL members to see what the final outcomes of these considerations were, we do know from public reports that at least a handful of libraries have decided to cancel their bundled arrangements with ScienceDirect.(5) This represents about 23% of those libraries that indicated they were considering cancellations this year. While the "Big Deal" is still a good deal for many institutions, it will be interesting to track developments over the next several years.

Canceling Print

One method libraries are using to control costs is to cancel print subscriptions where both print and electronic versions exist. In the fall 2002 survey, only a few libraries indicated that they had moved to electronic-only versions of the titles offered by these 14 publishers. In the more general survey conducted in 2003, many more libraries indicated they were making the switch. When subscribing to new titles or packages, 25 libraries (44%) indicated they were subscribing only to the electronic version. Another six indicated that while this wasn't yet standard practice, they seemed to be moving in that direction. Forty-three libraries (75%) indicated they were canceling print when they also subscribe to the electronic editions, although many qualified this answer to say this was still being done selectively. Given budget situations, this is a trend that is likely to continue over the next several years.

Usage Terms

The 2002 survey asked a number of questions regarding uses explicitly permitted or prohibited by the license agreement a library has with a publisher. Uses covered were interlibrary loan (ILL), electronic reserves, course packs, distance education, and remote access. While we asked respondents to identify explicit clauses, we do recognize that specific uses need not be explicitly addressed to be permissible. Some attorneys will advise that if a license is silent on an issue, one should turn to the appropriate law that would govern the relevant activity. For example, if a license were silent on ILL, a library should look to Copyright Law for guidance. In such an instance, a library could conclude that the use of electronic resources for ILL would be entirely appropriate and permitted. Another attorney, however, may take the position that unless a use is explicitly included in a license, it is not permitted. A library's position on this issue may reflect the judgment of its institution's legal counsel.

Based on the responses to the survey, most of the publishers, with the exception of Academic Press, Kluwer, and AIP, explicitly allowed many, if not most, of the libraries to use an electronic resource for ILL. As best we can tell, that use was primarily to allow libraries to print from the electronic copy and deliver the article through mail, fax, or Ariel.(6) At the same time, a number of libraries reported that using the electronic publications of these same publishers for ILL was prohibited. In those few cases where ILL is prohibited, the library may be working under an earlier version of the license of a publisher who has since changed its policy. Those cases where there is more of a mix of responses suggest that some libraries have been more successful in their negotiations. The data is unclear as to whether any of the libraries were allowed to transmit electronic copies of articles for ILL purposes. In the 2003 survey, 19 libraries identified 43 publishers with whom they have been able to negotiate ILL by distributing an electronic file. These publishers included 11 of the 14 covered in this paper, with the exception of Academic, Elsevier, and AIP. The number of libraries successfully negotiating this use with any one publisher ranged from one to five.

The use of articles from electronic resources was permitted by all 14 publishers for at least some of the libraries for electronic reserves, course packs, and distance education (see Table 6). It would appear that electronic reserves was standard or readily negotiated for Project Muse, ACM, Academic Press, Blackwell, and Wiley; course packs were permitted most frequently by Academic, Blackwell, Kluwer, and Wiley; and distance ed was permitted most frequently by Project Muse and Springer. At the same time, all of these publishers, except Project Muse and ACM, explicitly prohibited use of the licensed content in e-reserves, course packs and distance education (or required permission) for some libraries. Libraries have appeared most successful at ensuring remote access for their users. All of the 14 publishers overwhelmingly explicitly permit remote access, which corresponds with the responses from the summer 2003 survey where libraries identified remote access as one of the key make-or-break issues they have for license negotiations.

Bottom-Line Positions

Libraries were asked whether they have bottom-line positions that will make-or-break a deal. Eighty-six percent (48) of the respondents to this question said "yes" and identified over 32 different items. The most frequently mentioned were:

Governing law and venue (19) Indemnifications (18) Remote access (11) Price (11) Access by Walk-in Users (10) ILL (9) Fair Use (8) Archival Access/Perpetual Rights/"Own" the Content (7) Adequate Definition of Authorized User (7) IP Access (no username and password) (7) Definition of University/Campus as Single Site (6) "Escape Clause" for Multi-Year Contracts (6)

Non-Disclosure Clauses

From time to time there is a call within the community to encourage libraries not to accept non-disclosure clauses in their license agreements because these clauses prohibit subscribers from comparing terms to determine if they actually received the best deal available. Some states have laws prohibiting public institutions from accepting non-disclosure clauses. To find out more about this, ARL asked public institutions whether or not these agreements are legal in their state or province. Sixty-five percent (26) of the public institutions responding indicated that such clauses were legal. Of these 26 institutions, 20 reported that they did accept non-disclosure clauses in their licenses, although six libraries noted that these agreements were still subject to sunshine laws and FOIA requests. Two respondents noted that they would only agree not to disclose pricing terms. Of the 11 private institutions responding, nine (82%) indicated they accepted non-disclosure clauses.

One respondent clearly stated the dilemma of the discussion of non-disclosure:

Non-disclosure is a problem. It is certainly a method publishers use to defeat consortia, and to encourage libraries to compete with each other. It would be excellent if the ARL community could agree not to accept non-disclosure statements and yet, if the publisher offers to charge less in return for non-disclosure, it is difficult for libraries to resist; each of us is responsible for getting the very best deal we can for our own institution. Observations There are a few clear messages that arise from this analysis. Gathering publisher-specific information, let alone title-specific data, on journals, whether in print or electronic form, remains a challenge for libraries. The online order and receipt records are more likely to track current vendor information, rather than the publisher. Catalogers rarely have the time to keep publisher information up-to-date in bibliographic records unless there is a title change or another significant change requiring re-cataloging. Licensing information may be recorded, but it is often kept in a separate database. Efforts are underway under the auspices of the Digital Library Federation (DLF) to define data elements that should be included in library management systems to allow easier management and tracking of electronic resources. (7), When implemented, such systems should not only help local institutions enormously, but should allow libraries to respond to requests for information far more readily.

The study findings also suggest that libraries need to be tougher when negotiating uses of the resources they need. Some libraries have been successful at negotiating for electronic ILL, electronic reserves, course packs, and distance learning. Challenge a publisher who claims that no one has asked for a particular use before; chances are very good that someone has. The data presented here should provide some grounds for a solid counterargument. If a subscription is up for renewal and the terms have not been negotiated for several years, it may be time to review the contract. Look at the license the publisher is currently using; it may have better terms than your original agreement.

Most libraries considering canceling journal packages involve their faculty and administrators in some aspect of the decision-making process. Support from others on campus has been critical in the ability of libraries to either cancel journal bundles and/or negotiate better terms with publishers. Publishers seem not to have anticipated the willingness of faculty and administrators to support library actions.

A final observation concerns the reportedly rapid pace with which libraries are choosing electronic formats and canceling print. While most trust that the issues surrounding the preservation of digital files will be resolved, the fact that there will be no physical copies of these electronic resources leaves the library community vulnerable. No clause in a license guaranteeing perpetual access or any other user rights will help if the resource suddenly disappears for no matter what reason. This suggests the need for the judicious canceling and discarding of print and the urgency of projects to create distributed print repositories, such as that being spearheaded by the Center for Research Libraries. (8),


The two surveys provide a benchmark for helping describe the character and nature of the electronic environment in 2002/2003. The value and meaning of this information will increase with future research that will help us better understand this environment and identify trends. Since fall 2002, a number of changes have already taken place among the seven commercial publishers included in this report: Academic Press is now part of Elsevier Science, Kluwer and Springer have merged, and Taylor and Francis continues to acquire numerous other companies; publishers are continuously re-examining their pricing models and license terms based on the needs of their market; and many libraries are canceling print, while several have successfully challenged the "Big Deal." What will this environment look like in a few years? How will it differ from the benchmarks we have described in 2002/03? Stay tuned!

--Copyright 2004 Mary M. Case. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. To view a copy of this license, visit

This paper was presented at The New Challenge for Research Libraries: Collection Management and Strategic Access to Digital Resources, a conference sponsored by the University of Oklahoma, March 4-5, 2004, and is in press for the Journal of Library Administration. At that time the author was Director of ARL's Office of Scholarly Communication. ARL publishes this paper with permission of the author.


1) Data in this paragraph based on: Mark Young and Martha Kyrillidou, comps. and eds., ARL Supplementary Statistics 2001-02 (Washington, D.C.: Association of Research Libraries, 2003); and Martha Kyrillidou and Mark Young, comps. and eds., ARL Statistics 2001-02 (Washington, D.C., Association of Research Libraries, 2003).

2) Mark J. McCabe, "The Impact of Publisher Mergers on Journal Prices: A Preliminary Report," ARL 200 (October 1998): 5,

3) See: United Kingdom, Competition Commission, "Reed Elsevier PLC and Harcourt General, Inc.: A Report on the Proposed Merger," July 2001,; European Commission, "Case No. COMP/M.3197-Candover/Cinven/Bertlesmann-Springer: Regulation (EEC) No. 4064/89 Merger Procedure," July 29, 2003,; and Office of Fair Trading, "The Market for Scientific, Technical and Medical Journals: A Statement by the OFT," OFT396, September 2002,

4) In analyzing survey responses for the 2002 survey, there were occasions where individual answers were both inconsistent and unexplained. These answers were eliminated for the specific category; averages were determined based on the remaining valid answers for that question.

5) Charles Goldsmith, "Reed Elsevier Feels Resistance to Web Pricing," Wall Street Journal (Jan.19, 2004), B1+.

6) The responses on ILL in the 2002 survey are unreliable because the survey was not clear enough in its definition of electronic ILL. We had hoped to be able to distinguish between the right to print and then send a fax electronically through a service such as Ariel, and the right to download and send an electronic file directly. The definition in the 2003 survey was more explicit.

7) See Adam Chandler and Tim Jewell, A Web Hub for Developing Administrative Metadata for Electronic Resource Management (Ithaca, NY: Cornell University Library),

8) For more information on this project, see Preserving America's Print Resources (Chicago: Center for Research Libraries),

Case, Mary M. "A Snapshot in Time: ARL Libraries and Electronic Journal Resources." ARL, no. 235 (August 2004): 1-10.