Contact Us | Members Only | Site Map

Association of Research Libraries (ARL®)

  Resources Contact:
Lee Anne George
ARL: A Bimonthly Report
ARL: A Bimonthly Report, no. 234 (June 2004)

Serials Trends Reflected in the ARL Statistics 2002-03

Share Share   Print

by Martha Kyrillidou, Director, ARL Statistics & Measurement Program

Serial prices have been at the center of hot debate over the need for change in scholarly communication for the past two decades. Serial unit costs have increased 215% over the past 17 years according to the ARL Statistics 2002-03.1

The Numbers

In 1986, the average cost of a serial subscription for ARL libraries was $89.77; this cost increased to $283 by 2003 (see accompanying table). During the same timeframe, unit costs for books/monographs rose 82%, from an average of $28 to $52; library expenditures increased 128%; and the Consumer Price Index rose 68%.

As dramatic as these cumulative increases are, we should not overlook the fact that the average annual growth rate for serial unit costs since 1986 has fallen from a high of 10.2% in 1995 to 7% as of 2003. In the past few years, research libraries appear to have gotten more for their serials money than they did during the 1990s, perhaps as a result of canceling their most expensive journal subscriptions as well as purchasing the same content in dual format (print and digital) for an incremental surcharge.

The change in serial expenditures has been relatively level compared to the change in serial unit costs, although the growth rate of expenditures has slowed somewhat since 1995, from 8.6% to 7.8% in 2003. On average, ARL libraries spent $1.5 million on serials in 1986; they are currently spending more than $5.3 million. The cumulative increase in serials expenditures from 1986 to 2003 was 260%. Monograph expenditures increased 66% during the same time period. Libraries spent on average $1.1 million on monographs in 1986; they are currently spending $1.8 million.

Research libraries on average have been buying more serials during the past couple of years than they bought in any year since 1986. Whereas the average number of serials purchased in 1986 was 15,919, this number increased to 17,673 in 2002 and 18,142 in 2003. Also, the number of monographs purchased has rebounded to its 1986 level--about 32,600 books on average--for the first time over this period.

Nonpurchased serial subscriptions make up a growing amount of the content that libraries offer, increasing by an annual average of 6% since 1986. This category consists of a number of types of serials, including government documents, electronic serials made available free of charge with the purchase of print counterparts, and open access journals. The number of nonpurchased serials received by the average ARL library increased from 3,319 in 1986 to 8,873 in 2003.

Libraries are also providing access to more content through interlibrary borrowing and lending. On average, research libraries borrowed 22,146 items last year and lent 33,421 items. In 2002-03, they lent twice as many items as they did in 1986 and borrowed three times as many. These services were provided for roughly the same number of faculty and almost twice as many graduate students as compared to 1986.

Beyond the Numbers

It may be useful to consider these trends within their larger context, which is characterized by the proliferation of electronic serials and economic changes, from a booming economy in the '90s to tougher economic conditions more recently.

During the past five years, libraries have expanded the amount of material to which they provide access by purchasing the same content in new formats and acquiring new content, often through bundling arrangements, as well as by managing the growing amount of content available through open-access mechanisms. The purchase of new and dual-format content via bundling or "big deal" arrangements2 is probably partly responsible for the recent decline in the growth rate for serial unit costs--libraries have added serial titles to their collections at lower incremental prices. These additional titles are often duplicate subscriptions or titles the library would not otherwise purchase. Depending on the publisher's financial model, some of the additional content may be purchased or some may come bundled or "free" with a subscription to other products.3 Print was traditionally seen as the primary mode of dissemination in the '90s, with electronic as the secondary mode, but there are signs that authors, publishers, and libraries are ready to experiment with and embrace new models that reverse the balance, making electronic primary and print secondary.

We are still in the early stages of exploring the full potential of born-digital products and services.4 The bundled packages offered by publishers are not necessarily meeting the changing, high expectations of research library users. Greater awareness of the serials crisis has resulted in faculty and students supporting title-by-title cancellations instead of the big deal. Faculty and students are helping libraries improve their position at the negotiation table with commercial publishers of the scholarly record. And the Create Change campaign,5 SPARC, and the open access6 and alternative publishing movements are injecting competition into the serials marketplace.

In the past six months, a number of universities have taken action in support of libraries' decisions to withdraw from the big deal. In December 2003, the North Carolina State University Faculty Senate passed a resolution supporting the libraries' prerogative to "decline highly restrictive offers, such as those recently proposed by Reed Elsevier for its ScienceDirect online product."7 Also in December, Cornell University's Faculty Senate passed a resolution in support of the library's decision to forgo renewing the big deal.8 The Cornell resolution points out that Cornell libraries are buying 930 Elsevier titles [that] represent fewer than 2% of the total number of serials titles to which Cornell subscribes; the $1.7 million [the library spends on these titles] comprises something over 20% of the library's total serials expenditures, including those of the Medical School." Similar resolutions have been passed by University of California, Harvard University, Triangle Research Libraries Network,9 University of Connecticut, University of Maryland, Massachusetts Institute of Technology, and Stanford University.10

In summary, the slowing growth rate for serial unit costs from a peak of 10.2% in 1995 to 7% in 2003, may offer a glimmer of hope that eventually we will be able to contain price increases at a level closer to the general inflation rate of 3%. The slower growth of serials unit costs may be evidence that the academic community is beginning to behave like an informed consumer, looking for good deals that are sustainable and supporting the tailoring of subscription packages to increase value for money.


  1. Martha Kyrillidou and Mark Young, ARL Statistics 2002-03 (Washington, DC: Association of Research Libraries, forthcoming in 2004).

  2. Kenneth Frazier, "The Librarians' Dilemma: Contemplating the Costs of the 'Big Deal,'" D-Lib Magazine 7, no. 3 (March 2001),

  3. In the ARL Statistics, nonpurchased serials are not included in the calculation of serial unit cost.

  4. Electronic Scientific, Technical, and Medical Journal Publishing and Its Implications: Report of a Symposium (Washington, DC: National Academies Press, 2004).

  5. Create Change is an advocacy and education campaign cosponsored by ARL and the Association of College and Research Libraries to engage the academic community in reclaiming scholarly communication

  6. Richard K. Johnson, "Open Access: Unlocking the Value of Scientific Research," paper presented at "The New Challenge for Research Libraries: Collection Management and Strategic Access to Digital Resources," University of Oklahoma, March 4-5, 2004 (Washington, DC: SPARC, 2004),

  7. North Carolina State University Faculty Senate, "Resolution on Bundled Content and Elsevier" (Raleigh, NC: NCSU, December 2, 2003),

  8. Cornell Faculty Senate, "Resolution Regarding the University Library's Policies on Serials Acquisitions, with Special Reference to Negotiations with Elsevier" (Ithaca, NY: Cornell University Library, December 17, 2003),

  9. "Libraries Work with Faculty to Cancel Elsevier Titles," SPARC E-News (December 2003-January 2004),

  10. "Update: Library-Faculty Collaboration to Cancel Elsevier Titles," SPARC E-News (February-March 2004),

To cite this article

Martha Kyrillidou, "Serials Trends Reflected in the ARL Statistics 2002-03," ARL, no. 234 (June 2004): 14-15,