The State of Large-Publisher Bundles in 2012

F or well over a decade, research libraries have been spending millions of dollars per year licensing collections of journals published by just a handful of publishers. Ten years ago ARL surveyed its membership about their licensed collections of journal titles. In 2002, ARL asked for information regarding members’ expenditures for 60 journal publishers, ultimately reporting findings for the 7 most commonly subscribed publishers. In 2003, a second survey added further information about some licensing terms.2 ARL surveyed its members again in 2005 about their 2006 licenses with the 6 largest publishers at that time.3 Early in the summer of 2012, ARL again surveyed its member libraries about their subscriptions to journal collections from large publishers.4 The data collected in this most recent survey show that a great deal has changed in the last decade, and yet several issues remain concerns within the library community. Pricing models and license terms, consortial arrangements, and the conversion from print to electronic subscriptions remain issues across the surveys.


Signs of Change
Comparing surveys from 2002, 2005, and 2012 shows the steady uptake of the largest publisher bundles.Three of the four publishers covered in the two most recent surveys (Elsevier, Springer, and Wiley) are now licensed as bundles by 90% or more of libraries for which data were collected (see Table 1).Even for Taylor & Francis, which was the least frequently licensed bundle in both surveys, market penetration tripled.Although several libraries have discussed dropping bundles, it appears within this market segment that, overall, subscription has grown over the last six years.Some libraries may have dropped bundles but, if so, their actions were more than counterbalanced by libraries taking on licenses for journal bundles.For 2012, the average responding libraries had licensed more than five bundles among the seven publishers.However, the definition of the subscription bundle has changed for many libraries since the 2006 subscription year.There is evidence of substantial pruning of journal bundles through the "Great Recession."Six years ago, many more libraries described their bundles as including all of a publisher's titles.For the four publishers covered in both studies, percentages of libraries reporting the full publisher list of titles dropped noticeably for all (see Table 2).The data are consonant with a shift in consortial licensing from packages of publishers' full lists to pooled lists of titles licensed by consortia members.However, for those publishers in both surveys, substantially more libraries are now defining their bundle at the institutional level.In 2006, fewer than 10% of subscribing libraries defined their own bundles with Wiley, Springer, and Taylor & Francis (rather than as a consortial pool or the full list).By 2012, the percentage of libraries with institutionally defined bundles tripled for Wiley and Taylor & Francis licensors, while the proportion institutionally defining the Elsevier bundle more than doubled.Even the proportion of institutionally defined Springer licensors increased somewhat.Over the last decade, ARL's member surveys have documented a remarkable shift in the move of research libraries away from collecting current journals in print form.Already in 2003, Case was observing a rapid shift from bundled print and electronic subscriptions to electronic-only subscriptions. 6By 2006, print equivalents had been cancelled for 153 out of 266 contracts (58%) for the five publishers covered by that survey.In 2012, with the exception of a single publisher (Taylor & Francis), no libraries reported that they retained corresponding print subscriptions for their complete journal bundles (7 out of 352 contracts still included print, or 2%).However, 23% to 43% of respondents (by publisher) indicated that they still carried some selected print titles in 2012.This suggests that there are some instances where publishers have not yet fully implemented digital publishing or developed fully substitutable digital versions for some journals; or, perhaps some users are not yet ready to support cancellation for individual titles.
The ability to cancel titles (or not) has been a concern about journal bundles from the beginning.The most recent survey reinforces the findings from the earlier ARL surveys that only a small minority of libraries retain any ability to cancel titles within their bundle during the term of their license (for those able to disclose, 39% of contracts forbid any cancellations under any circumstances during the length of the contract, and another 24% cannot do so without declaring a fiscal emergency).However, the reduction in full-list bundles reported suggests that, between 2006 and 2012, many libraries have reduced the extent of their bundles between contracts.Given the budget reductions reported in ARL's surveys on library budgets conducted in 2009, 2010, and 2011, following the onset of the most recent recession, 7 one might anticipate that libraries would press for more liberal cancellation options.However, there is no evidence within the survey data to support this hypothesis.

How Long Can Historic Pricing Last?
To an outside observer it might be surprising to see that more than a decade into electronic journal licensing, historic subscription expenditure remains by far the dominant model for determining library pricing.For most of the 375 contracts for which pricing model was reported, libraries' prices are determined by their historic print subscriptions (now dating back to the 1990s in many cases), plus inflation increments applied annually during the intervening years.For the four largest publishers, 82% of contracts are priced in this way.This could be seen as even odder given that the survey also provides evidence of many institutions' having reduced the content in their bundles from the publisher's full list to some subset defined for a consortium or at the institutional level.Two publishers appear to have largely shifted their pricing models to either a form of tiered pricing (larger institutions pay more than smaller institutions) or simply an aggregate subscription price for all of the titles in the bundle.In contrast to the four largest publishers, the American Chemical Society and Nature differ not just in their abandonment of historical pricing strategies but as publishers with small title lists and control of top-tier journals in their field.While usage-based pricing has generated some discussion as a potential alternative pricing model for large-publisher collections, it does not appear to have found expression in any significant way within the research library community.

Licensing Terms and Clauses
The research library community shares an interest in most journal contract terms and clauses with other types of libraries.There are additional terms that are also of particular interest, because of the nature of a research library clientele.The 2012 survey included questions about some specific terms and clauses now being included in licensing agreements in order to understand the current licensing environment.
The ability to share information about contract terms, as well as pricing information-and thus allow the research library community to collectively advance its positions regarding access to content-is dependent on the knowledge of what is in the agreements.In May 2009, the ARL Board of Directors, based upon the recommendation of the Advancing Scholarly Communication Steering Committee, issued a statement encouraging members to refrain from signing nondisclosure clauses in their licenses.The 2012 survey asked two questions regarding nondisclosure.The first was whether members had policies to refrain from signing nondisclosure clauses.Of the 50 respondents, just under half (49%) indicated that they had such a policy now.Respondents' comments provided some useful perspective: for many institutions, even if the signed license states that nondisclosure is required, the institution abides by the law of their state, and respondents presumed that open records acts apply.
Recent attention surrounding nondisclosure clauses has, however, resulted in an increasing number of institutions that are now in the process of establishing a policy not to sign such clauses.An interesting comparison can be made to data collected for the 2003 survey.In that survey, ARL asked about the legality of licenses with nondisclosure clauses.Twenty-six public institutions reported that they were legal in their jurisdictions, and 20 (77%) of those 26 institutions did accept those clauses, although some reported that they were subject to sunshine laws and FOIA requests. 8A similar percentage of 11 private institution responses was reported.It is encouraging that over this ten-year period, more institutions are attempting to refrain from signing nondisclosure clauses.
The second question in the recent survey asked whether nondisclosure clauses were signed for the 2012 contract for specific publishers (Table 4).Depending on the publisher, a range of 10-28% of respondents indicated that they had indeed signed a nondisclosure clause.When comparing the data with that of 2006, it appears that significant progress was made with Wiley and Taylor & Francis, since the percent of nondisclosure dropped by two-thirds and by more than half, respectively.Elsevier decreased slightly, but Springer, however, rose 11%, raising questions as to whether it increased the pressure on libraries to sign nondisclosure after the open-records decision in the state of Washington in 2009. 9Again, the comments for 2012 indicated that some libraries assumed open records laws applied to contract terms.
The role of consortia cannot be overstated here, since they negotiate many large bundles on behalf of the libraries.Those consortia that have recently or are now establishing nondisclosure policies will affect the responses to this question in the future.ARL also is supporting model author-rights clauses that provide for open access deposit of works by the institution's authors.Only a very small number of libraries did attempt to negotiate this clause for 2012 contracts, but of those libraries there are success stories of at least 50% or more for nearly all of the publishers.This should encourage other research libraries to consider adding this clause to their license, and respondent comments indicated that some were expecting to do so in the future.
The ability to text mine is a very new license consideration, and libraries were asked if they were beginning to negotiate that clause.No conclusions could be drawn from the low response to this question, but comments indicated that many research libraries are beginning to work with the journal publishers to address this need.
In addition to nondisclosure and author-rights clauses, interlibrary loan terms in licenses-which vary among the large-publisher bundles-are also of particular interest to ARL.The 2012 survey asked libraries to identify any clauses that address fulfillment of ILL requests.The responses included information on publishers that allow sending printed articles (58% to 79%), allow the transmission of electronic articles (39% to 73%), and allow international interlibrary loan (11% to 36%).A small number of libraries indicated that the contract was silent on ILL.The variation in response, and a review of the comments, indicate a lack of consistency among publishers and even among libraries that may have signed agreements with the same publisher.The challenge for future contracts is to make sure that any licenses being signed do not abridge a library's ability to share materials.
Two other questions about general license terms dealt with preservation, and the responses indicated that for each publisher slightly more than half of the 2012 contracts addressed preservation of licensed content either locally or through third parties, though the specifics of how that will be done were not requested in the survey.More specific surveys will need to be done to determine how the large packages are being preserved and the license clause language used to ensure that long-term preservation needs are being met.
The final question about specific terms asked whether access could be retained for content that was subscribed to in the past.High percentages were reported for each publisher (76% to 90%), but comments indicated that some publishers require a fee.License terms that provide research libraries with the ability to provide access to content they have previously licensed is a potential area in which to advance new and more detailed terms.
What Is Next?
The information about ARL member libraries collected for this study demonstrates ongoing strains in libraries' relationships with publishers and in their ability to maintain electronic journal bundles in difficult financial times.A great divide is evident between the anticipated promise of electronic publishing foreseen at the birth of journal bundles and the reality of large-publisher journal licensing in 2012.Fully digital collections of big publishers' journals are the norm in research libraries, and license negotiations are beginning to address issues such as text-mining rights, harbingers of the potential benefit of a fully digital scholarly communication system.
Yet, the expected opportunity to make more content available to libraries than could be afforded through print subscriptions seems to have largely evaporated.Research libraries are more frequently subscribing to subsets of publishers' lists than they were six years ago, while still paying for their access largely based on print subscription costs grounded in an increasingly distant past.Content and pricing seem to be trending toward a growing disconnect.Consortial licensing offers continuing benefits to institutions, but less and less often that benefit seems to be including access to a larger body of content.
The licenses that are signed by the libraries or the consortia that represent them need stronger language in the clauses that are important to the research library community.The ability to use the content in new ways, the importance of sharing information between and among researchers and institutions, and the expectation that content will be available in the future all require diligent examination of and development of clauses that meet those needs.Especially important will be the emphasis on moving away from nondisclosure clauses to better position research libraries to negotiate license terms and prices individually or with their consortia more effectively.
The authors acknowledge the value of the time taken by respondents to answer the survey and thank them for their willingness to share information about their contracts. Endnotes

Table 1 .
Research Library uptake of Journal bundles * Data based on combination of responses to ARL 2005 survey of "have licensed [for 2006]" and "planning to license [for 2006]."

Table 2 .
Research Library Definition of Journal Bundles

Table 3 .
% of contracts Licensed via consortium

Table 4 .
Libraries that Reported signing nondisclosure clauses