With the economy the way it is has been over the past year or so a lot of libraries have turned back to ILL as an alternative to shrinking book-buying budgets. I wonder if this is really working the way we want it to.
In the most recent study I could find about the cost for mediated ILL transactions the Association of Research Libraries showed that the average cost of ILL is about $27.00 USD. $17 for the borrowing library and $10 for the lending library. Even if we eliminate the lending library costs, $17 could purchase a fair amount of titles, especially popular but not necessarily brand-new ones.
If the study done in 2002 on patron-driven collection development can be duplicated in other libraries, 68% of titles requested through ILL which were purchased received at least one more use (see Anderson, et al). In this study, the librarians limited their pilot to scholarly non-fiction titles (Anderson 10). Presumably popular reading titles would see even more usage which could be of particular interest to public libraries who are active ILL participants.
There can be even more benefit to ILL in cases where the borrowing library is requesting a title that they own but all copies are checked out. ILL and holds/recall operations really should work together. Some colleagues of mine and I have been looking at the holds queues in relation to ILL requests. In many instances it is more cost-effective to purchase another copy of a title to offset existing holds rather than spend money on ILL for the title. These kinds of items usually have a large usage potential automatically as is evident by the extended holds queues for local copies. Why spend $27 for one use when you could spend a comparable amount and get 10s if not 100s of uses?
If an item costs the same as an ILL request, purchasing it will at the very least be the same cost-per-use as ILL. But if it circulates even one more time, the borrowing library copy now effectively is 25% of the cost-per-use of ILL. Ordering the book twice through ILL would be almost $60 but if it was purchased once ($30), then used again, the cost-per-use drops to $15. For really popular items the holds queue essentially guarantees several uses, which would bring the cost-per-use down to mere cents instead of dollars.
If a library is not worried about shelf space it makes more sense economically to not ILL items which are popular–purchasing another copy or two for local users is potentially much cheaper in the long run.