As of 22 April 2009 this website is 'frozen' in time — see the current IFLA websites

This old website and all of its content will stay on as archive – http://archive.ifla.org

IFLANET home - International Federation of Library Associations and InstitutionsAnnual ConferenceSearchContacts

62nd IFLA General Conference - Conference Proceedings - August 25-31, 1996

Determining the Cost of An Interlibrary Loan in North American Research Libraries: Initial Study

Joan Chambers,
Dean of Libraries,
Colorado State University


(Based on a talk given at the New Mexico Library Association 70th Annual Conference)

In 1992 the Association of Research Libraries (ARL) and the Research Libraries Group (RLG) collaborated on a study to determine the cost of an interlibrary transaction. The purpose of the study was to provide benchmark data that research libraries could use as a management tool to inform decisions about whether to buy or to borrow materials. This paper summarizes the conditions leading to and the results of the ARL/RLG ILL Cost Study.


I. What Prompted the Cost Study?

By 1991, numerous factors had converged to cause research libraries to focus on their interlibrary loan operations, culminating in the most definitive interlibrary loan cost survey undertaken up to that time. Many articles had already been written about the explosion in the amount of published information--doubling every seven to ten years or less, depending on who was reporting. According to The Bowker Annual, 1992, the number of academic titles published in North America increased from just under 73,000 in 1985/86 to over 103,000 in 1990/91, a 41% increase (1). Over the same five years, the median number of titles purchased by Association of Research Libraries (ARL) declined from just over 32,000 to slightly over 29,000, a 9% decline which widened the gap to be covered by interli brary loan (2). The same pattern was true for the gap between the number of current serials published as compared to the median number purchased.At the same time, the scope of material that had come under bibliographic control was enormous and growing, thanks primarily to the ability to generate machine readable records and distribute them as online catalogs and CD-ROM databases. This resulted in increased awareness of and requests for materials. In a 1991 article in the Journal of Interlibrary Loan & Information Supply, Suzanne D. Gyeszly and Gary O. Allen reported the results of a study done at Texas A&M University (3). The study concluded that the change from paper to online indexes increased the number of interlibrary loan users as well as the number of requests. The study con firmed the increase in accessibility of materials through online indexes and emphasized the increased importance of interlibrary loan services in bridging the gap between access and availability.

Personal computers, campus networks and the Internet were also changing users' attitudes and expectations about access to information, availability of materials and the speed with which they wanted them to be delivered.The spiraling costs of library materials, especially of foreign journal subscriptions in science and technology, were well documented. For example, at Colorado State University th e average cost of a serial subscription increased from $134 in 1986/87 to $291 in 1991/92, an increase of 117%. (The median increase in ARL libraries was 72%.) During that same five years, while total serial expenditures in ARL libraries increased 70%, the number of serial titles purchased declined 2%. A similar pattern held for monographs. While costs increased 47%, expenditures went up only 15% and the number of volumes purchased went down a staggering 15% (4) (5). In ARL institutions, library expenditures as a proportion of university education and general expenditures had declined from a high of 3.9% in 1979 to 3% in 1990. This decline had contributed to the difficulty libraries were having in keeping up with the explosion in the amount of published information and the spirali ng cost of library books and journals (6) (7).In October 1992, a Quick SPEC survey conducted by ARL revealed that in 71 institutions that had targeted specific dollar amounts, planned journal cancellations averaged about $169,000 per library, for a combined total of as much as $12 million (8). This followed on the heels of announced cancellations totaling just under $7 million in 1991.

The survey also revealed that somewhere in the neighborhood of 139,103 titles had been canceled by ARL libraries since 1987. For example, since 1988, eight libraries in the Greater Midwest Research Library Consortium had canceled a combined 17,761 subscriptions, totaling $3,860,513. This unprecedented level of cancellations had increased their reliance on interlibrary loan.Since 1986, interlibr ary borrowing by ARL libraries had increased 47% and interlibrary lending had increased 45% (2). Staff had been hired or reassigned from other library services to absorb this staggering increase. New equipment, such as PC's, fax machines and scanners, had been installed to help handle the volume of requests. New and streamlined procedures had been implemented. Nevertheless, the increase was o utstripping the capacity of libraries to respond. For example, at Colorado State University interlibrary borrowing had shot up 171% since 1985/86, from 7,048 to 19,254 items. Lending had increased 77%, from 16,151 to 28,584 items.

II. Other Related Actions

Several other actions were related to the ILL Cost Study. In 1992 ARL libraries were surveyed on their lending and borrowing practices and policies, including reimbursement agreements and the use of commercial vendors for document delivery. The results of the survey were released as two ARL Spec Kits: #184, Interlibrary Loan Trends: Making Access a Reality; and #187, Interlibrary Loan Trends: Personnel and Staffing (9) (10). Based on their analysis, authors Virginia Steel and Tammy N. Dearie identified three areas that warranted further attention: ILL fees, the use of commercial document suppliers, and user-initiated requests.To assimilate and build on the survey findings, a white paper, Maximizing Access and Minimizing Cost: A First Step Toward the Information Access Future, was written by Shirley Baker and Mary Jackson for the ARL Committee on Access to Information Resources. The paper, which was revised in February 1993, included a critique of current interlibrary loan processes and then proposed an ideal system (11). In response to the white paper, the Committee recommended two courses of action: a 3-year pilot project with OCLC to design the "ideal" ILL su bsystem and a project with the Center for Research Libraries to develop an economic model for document delivery.

At the same time, the 1980 national interlibrary loan code was being revised to be less prescriptive and more encouraging of resource sharing (12).Finally, OCLC had announced its intent to create a comprehensive information database on library costs, workflows and usage. As a first step, the project would collect, review and synthesize data on costs, workflows and usage in the general domains of ILL and document delivery (13).The time was obviously right for the definitive ILL Cost Study which was jointly undertaken in 1992 by the Association of Research Libraries and the Research Libraries Group, with the support of a supplementary grant from the Council on Library Resources. The purpose of the study was to obtain benchmark unit borrowing and lending cost data for interlibrary loan which would enable library managers to make comparisons between ILL and alternative methods of information delivery. The study focused exclusively on costs directly associated with interlibrary borrowing and lending. The methodology was adapted from the cost study model published by Dickson and Boucher in Research Access Through New Technology (14).

III. The Survey

The study was based on an eleven page survey, which was accompanied by nine pages of instruction in an effort to ensure both comprehensive and comparable data. Following the general profile of the responding interlibrary loan unit, data was collected for both borrowing and lending in the following categories of costs: supervisory and non-supervisory staff; network and communications; delivery; photocopy; supplies; equipment and software; rental, maintenance and annual software licensing; and fees for purchasing from other suppliers. Income which offset costs was also identified.

IV. Survey Results

Seventy-six institutions completed the survey. Each received the results in a set of tables which gave the individual institutional data only for the responding library. All other data was reported in the aggregate and no data was made available for comparison on a library-by-library basis. Six pages of tables included the individual library data in each of the eight cost categories for both b orrowing and lending, as well as aggregate data for all libraries with breakdown by mean, median, and percentile (10, 25, 75 and 90). Scatter diagrams illustrated the results for both borrowing and lending for unit cost by type of library (Canadian or U.S., public or private), unit cost by geographic region (North, Midwest, South, West, Canada), unit cost per filled transaction, staff cost by vo lume or level of activity, staff cost by geographic region, and unit cost by photocopy activity.

When data from all participating libraries were aggregated, key findings included the median unit cost for a research library to lend an item was $9 and the median unit cost to borrow an item was $18. For 80% of the libraries, the unit costs ranged from $6 to $17 for lending and from $10 to $30 for borrowing. The largest cost component was staff salaries. In 1991/92 a total of 4.1 million item s were loaned by and 1.4 million items were borrowed by ARL libraries (15). Extrapolating the results of the cost survey suggests that in 1991/92 ARL libraries spent over $71 million on interlibrary loan operations -- $26 million for borrowing and $45 million for lending.It was not likely that interlibrary loan departments could continue to absorb the unrelenting demand for services or that libr aries could continue to increase expenditures to support the growing demand. For many transactions, turnaround time did not meet the needs of the end-user. And finally, libraries that were net lenders could not continue to subsidize through ILL services libraries that were net borrowers.

V. New Models for Document Delivery

The results of the study caused librarians to begin to question the status quo. Was there a new model for delivering information to users that was less labor intensive than mediated interlibrary loan and that could reduce unit costs, reduce the rate of increase in ILL volume, reduce the impact on net lender libraries and improve turnaround time for end-users? Could the volume of ILL activity be held constant at 1991 levels? Could future increases be absorbed by enabling users to bypass ILL staff for some of their document delivery needs? The same factors that converged to cause research libraries to focus on their interlibrary loan operations also contributed to the expanding development and marketing of services designed to provide speedy delivery of documents to end-users. Prices for commercial document delivery services were clearly competitive with the range of costs found in the survey, especially when one considers that for one interlibrary transaction the median cost was $18 for the borrowing library plus $9 for the lending library for a total cost of $27 (16) (17) (18). Direct end-user request and delivery is less labor intensive than mediated document delivery bec ause it bypasses interlibrary loan staff altogether. It does not increase current levels of ILL volume. It does not impact another library that may already be a net lender. It improves turnaround time by eliminating the middle man--the ILL unit--and delivers the document directly to the user's computer or fax machine or mails a photocopy. It was becoming increasingly clear that self-service de livery could be an integral part of the whole document delivery system, complementing the many technological advances that had been incorporated into mediated interlibrary loan services--use of telefacsimile; scanning and digital transmission (such as RLG Ariel); electronic transfer of requests to ILL; and the invisible pass-through of ILL requests to commercial suppliers, many of which were avai lable on the RLIN and OCLC ILL subsystems. Many ILL units were already buying documents from commercial suppliers, in addition to borrowing from other libraries. A three-pronged approach to delivering the information needed by users was evolving:

While perfect solutions were not at hand, workable experiments and pilot projects were in order. Interlibrary loan departments had demonstrated leadership, initiative and ingenuity in streamlining, automating and enhancing ILL services in order to absorb staggering increases in requests and to improve turnaround time. Many of these units had already moved beyond interlibrary loan services to be come mediated document delivery services. These advances were ably described by Julie Wessling in "Document Delivery: A Primary Service for the Nineties" (19). In "Document Delivery", Malcolm Getz made a compelling case for integrated document-delivery services, which incorporated two of the above approaches: mediated document delivery from remote providers, which he referred to as "external" and delivery from onsite sources, which he referred to as in-house, on-premises or internal fulfillment” (20).

VI. Challenges Ahead

Many challenges remain before a total integrated document delivery system becomes a reality. Cancellations and decreased purchasing power are causing research collections to look more and more alike as libraries are reduced to buying essential materials that support core programs. Vendors are in business for a profit and what they make available for document delivery will be driven by the marke tplace, not by the needs of individual scholars for esoteric, rare or foreign publications. A critical challenge for research libraries is to act collectively in a comprehensive and systematic manner to prevent the gaps of today from resulting in permanent loss of access for future generations of scholars and to ensure that unique literary manuscripts and archives--the primary materials that are the basis of research and scholarship--continue to be collected. This collective responsibility for our published heritage encompasses not just ownership, but also bibliographic control, preservation, archiving and delivery.While gateways and front-end user interfaces are making it easier for end-users to navigate the Internet and do their own searching, many practices will have to be standardi zed if self-service is to become commonplace. Standards will make it easier for users to gain a reasonable level of competence in searching with a minimal level of training.Currently libraries are funding mediated delivery and onsite collections. The ARL/RLG ILL Cost Study demonstrates the significant cost of mediated delivery for both the borrowing and the lending libraries. The ARL statisti cs have also tracked the unit costs for journals and books, which currently exceed $211 and $45 respectively (21). If direct end-user services are to become an integral part of the document delivery system, each library will have to grapple with the questions of funding and equity of access. Will costs be transferred to or shared by individuals, departments or colleges? Will deposit accounts b e used to cap services? Will library resources that currently support mediated delivery and onsite collections be reallocated to support end-users? How can a workable degree of predictability be brought to the budgeting process? An integrated document delivery system cannot be separated from collection development policies. Which collections will be fully supported onsite? Which will be suppo rted through mediated document delivery? Which will be left to the responsibility of end-users and to what degree will these be subsidized? The ARL/RLG ILL Cost Study can help librarians answer these questions (22) (23). In the previously cited article, Malcolm Getz also presented a theoretical model for determining what he called the "tipping point" in deciding when the number of uses makes it more cost effective to own than to pay for external document delivery (20).Expanding the universe of information available and then delivering what is requested in a timely fashion, whether from onsite or remote sources, will continue to challenge our ingenuity to develop and experiment with new models.


  1. The Bowker Annual Library and Book Trade Almanac, 37th ed. New Providence, NJ: R.R. Bowker, 1992.

  2. ARL Statistics 1990-91. Washington, DC: Association of Research Libraries, 1992.

  3. Gyeszly, Suzanne D. and Gary O. Allen. Effects of Online Periodical Indexes on Interlibrary Loan Services and Collection Development, Journal of Interlibrary Loan & Information Supply, vol. 1, no. 3, 1991, pp. 39-48.

  4. Okerson, Ann and Kendon Stubbs. The Library "Doomsday Machine", Publishers Weekly, 238, February 8, 1991, pp. 36-37.

  5. Okerson, Ann and Kendon Stubbs. ARL Annual Statistics 1990-91: Remembrance of Things Past, Present...and Future?, Publishers Weekly, 239, July 27, 1992, pp. 22-23.

  6. DeLoughry, Thomas J., Research Libraries Said to Face an Era of Tumultuous Change, The Chronicle of Higher Education, February 17, 1993, p. A15.

  7. Cummings, Anthony M., et al. University Libraries and Scholarly Communication, Washington, DC: ARL, 1992

  8. Serial Cancellations Survey Shows Steepening Trends, ARL, 166, January 1993, p.10.

  9. Dearie, Tammy N. and Virginia Steel. SPEC Kit #184, Interlibrary Loan Trends: Making Access a Reality. Washington, DC: ARL, 1992.

  10. Dearie, Tammy N. and Virginia Steel. SPEC Kit #187, Interlibrary Loan Trends: Personnel and Staffing. Washington, DC: ARL, 1992.

  11. Baker, Shirley K. and Mary E. Jackson. Maximizing Access, Minimizing Cost: A First Step Toward the Information Access Future, Washington, DC: ARL, November 1992 (Revised February 1993).

  12. American Library Association. Reference and Adult Services Division. Management and Operation of Public Services Section. Interlibrary Loan Committee. National Interlibrary Loan Code for the United States 1993. RQ 33, no. 4, summer 1994, pp. 477-479.

  13. Prabha, Chandra. Costs, Workflows and Usage Data for Interlibrary Loan and Document Delivery, Annual Review of OCLC Research, July 1991-June 1992, pp. 18-21.

  14. Jackson, Mary E., ed. Research Access Through New Technology. NY: AMS Press, 1989.

  15. ARL Statistics 1991-92. Washington, DC: ARL, 1993.

  16. Ensor, Pat. Automating Document Delivery: A Conference Report, Computers in Libraries, vol.12, no. 11, December 1992, pp. 34-37.

  17. Leach, Ronald G. and Judith E. Tribble. Electronic Document Delivery: New Options for Libraries, Journal of Academic Librarianship, vol. 18, no. 6, January 1993, pp. 359-364.

  18. Khalil, Mounir. Document Delivery: A Better Option?, Library Journal, vol. 118, February 1, 1993, pp. 43-47.

  19. Wessling, Julie. Document Delivery: A Primary Service for the Nineties in Advances in Librarianship, vol. 16, 1992, pp. 1-31.

  20. Getz, Malcolm. Document Delivery, The Bottom Line, vol. 5, no. 4, winter 1991, pp. 40-44.

  21. ARL Statistics 1994-95. Washington, DC: ARL, 1996.

  22. Roche, Marilyn M. ARL/RLG Interlibrary Loan Cost Study. Washington, DC: ARL, 1993.

  23. Barrett, G. Jaia. The Cost of Interlibrary Loan, ARL, 166, January 1993, pp. 1-2.