Developing new funding models to support and sustain scholarly monograph dissemination in the humanities and social sciences is a topic of keen interest and critical importance. Key stakeholders include administrators of American colleges and universities and their tenure-track faculty, university presses, and the academic community at large. Pressure to address the problems inherent in the current system of scholarly publishing in the humanities has prompted several related research projects supported by grants from the Andrew W. Mellon Foundation.

One possible funding method that appears to be gaining traction is that of direct institutional author subventions for monographs, a model proposed in a prospectus issued by the Task Force on Scholarly Communication.[1] The Association of American Universities (AAU) and the Association of Research Libraries (ARL) created the Task Force in order to better examine and address the issues and challenges faced by the current system of scholarly monograph publishing in general, and specifically, “to address the inability of a market model to adequately support research monograph publication based primarily on scholarly merit. This prospectus describes a faculty title subvention designed to ensure the long-term economic viability of foundational scholarly monographic publishing, while promoting the emergence of innovative digital models.”[2] The prospectus is based on a white paper prepared for the Task Force by Raym Crow of the Chain Bridge Group.[3] These issues have also been discussed in the Wired Campus blog from The Chronicle of Higher Education.[4]

In order to study the viability of this institutional funding model in the context of real-world examples, the Andrew W. Mellon Foundation awarded a grant to Indiana University and the University of Michigan to “explore how a potential ‘subvention’ system of funding in which universities and colleges paid the costs of producing humanities monographs (rather than expecting publishers to recoup their costs from consumer payments) could work at two major research universities.”[5] The resulting monographs would be distributed via open access under a permanent nonexclusive license, thus greatly adding to their potential readership and ease of use. At the same time, the publishing press would be free to negotiate with the author to produce other versions of the work, including commercial versions.

Two of the main themes or concerns that arose during the course of the case studies can be thought of as a two-sided coin, with the “free riders” problem on one side, and an “unintended consequence” problem on the other. The “free rider” problem exists when works published by members of the Association of American University Presses (AAUP) are written by scholars at institutions that are not AAUP members and hence do not contribute to the financial support of monographic publications by university presses, almost all of which require financial support beyond that which they are able to obtain by selling their books.[6] Thus the current US publishing system for scholarly monographs relies in part on a relatively small number of university presses (and their institutions) that bear some of the costs of publishing works by scholars from other institutions without university presses, or by authors without an institutional affiliation. This financial burden to presses (and their current sources of subsidy) would be lessened under a direct subvention system, because the authors’ institutions would provide financial contributions to the university presses that are publishing their faculty’s works.[7] We note that much of the publishing cost (approximately 90 percent, per personal communication with Joseph Esposito)[8] is paid for in sales, and that only the portion of the publication cost that is subsidized from sources other than sales is subject to “free ride.”

Another concern that emerged in the course of conducting faculty focus groups and interviews with senior administrators at Indiana and Michigan was the flip side of the coin—that of the “unintended consequence of increasing inequality”[9]—namely, that a subvention system that seeks to democratize access on the part of users, supported by a subvention from the author’s institution or similar party, might not be financially feasible for some authors, exacerbating inequalities between “haves” and “have-nots.”

To our knowledge there are no existing studies that attempt to measure the extent of the “free riders” problem and the number of scholarly authors who could be negatively affected by a move to a direct subvention system. Such measurement requires knowledge of the institutional affiliations of the authors being published by American university presses. A preliminary examination of author affiliation data at Indiana University and the University of Michigan presses as part of the case studies suggested that the impact of a direct subvention model could be “significant,”[10] leading us to undertake the study reported on here.


We began with data from a grant-funded study conducted by Joseph Esposito and Karen Barch in which the goal was to compile a list of all primary monographs published by American university presses from 2009 to 2013.[11] Sixty-three of the potential sample of 106 presses successfully submitted data for that study, whose focus was on scholarly primary monographs in the humanities. Here, “scholarly monograph” is used not in the more narrow library-based sense of format (book form as opposed to a serial or microform, for instance). Joseph Esposito reports on the definition of “monograph” used in the Esposito-Barch study as follows: “We amended the John Thompson definition (see his "Books in the Digital Age"), which defines a monograph as a book by scholars for scholars—to which we added: and not including collections of essays. When the presses tagged their publications as monographs, that was the definition they were looking at.”[12] Another important distinction is that we looked at what were termed “primary” scholarly monographs, meaning works of original copyright and excluding revised editions or reprints.

Oxford University Press and Cambridge University Press are not included in the study. Those presses are, of course, much larger than any other members of the AAUP, and their homes, like those of some other members of AAUP, are abroad. The Esposito-Barch study excluded data from presses based abroad, and, perforce, we follow their lead.

Additionally, although we would have preferred to sample all publications, we were only able to obtain data from the study for a sample of presses. Thus, we have all publications from 2009 to 2013 for 15 presses, accounting for 2579 primary humanities monographs, or 24 percent of the total for the period as calculated by Esposito and Barch.

The sample of presses was chosen to mimic approximately (excluding OUP and CUP) the shares of works produced over the period by presses in each of AAUP Groups 1–4.[13]

The distribution of primary humanities monographs by Group was as follows:

In total, 10,689 total primary humanities monographs were published over the 5 year period by all Groups combined. Of those, 28 percent were published by Group 4 presses (2,992), 42 percent by Group 3 presses (4,474), 14 percent by Group 2 presses (1,449), and 17 percent by Group 1 presses (1,774) (101 percent due to rounding).

We approximated this distribution as best we could by examining 3 (of 9) participating presses in Group 4, accounting for 39 percent of the 2,579 titles that we examined; 3 (of 13) participating presses in Group 3, accounting for 32 percent; 4 (of 11) participating presses in Group 2, accounting for 17 percent; and 5 (of 24) participating presses in Group 1, accounting for 14 percent. Relative to the population, our share of works published by Group 4 presses is high (39 percent in the sample, 28 percent in the population) and our share published by Group 3 presses is low (32 percent in the sample, 42 percent in the population). Overall, we had data for 24 percent of the works and 26 percent of the presses in the Esposito-Barch study.

As we shall see, the shares of “free riders” and shares of authors by different Carnegie classifications differ noticeably by Group, justifying our decision to stratify our sampling by Group (given that we could not assemble a random sample of all publications).

Adding Author Information

For each of the 2,579 primary humanities monographs published by the 15 presses for which we had data, we sought to identify the institutional affiliation of the primary author (in some cases a volume editor), based on the data fields of title, publisher, primary author, and ISBN from the Esposito-Barch study. In order to record author institutional affiliation information, titles were searched on their respective press web sites. Most book descriptions on the press catalog pages include biographical information about the authors, and this served as the primary information source. In cases where biographical information was not listed, or further details were needed (such as the particular campus an author was affiliated with for those institutions with multiple campuses), further searching was done via, Google Books, Wikipedia, institutions’ websites, or authors’ web pages. The following information was recorded: primary author’s home institution; location of home institution (if one with multiple campuses); an indication of the author’s role, if known (e.g., editor, translator), and the author’s status or title if known (e.g., faculty member, professor emeritus , fellow, deceased, curator, and so on).[14] Information using these same data fields was recorded for up to two additional co-authors or editors. If an author or editor was not affiliated with any institution (for example, an independent researcher or writer, journalist, business person, military member, and the like), the home institution was designated as “no institutional affiliation.” There were 249 instances (9.7 percent of all works in our sample) where the author was not affiliated with an academic institution. When information about the author could not be located, “unable to locate” was entered. This means the institutional affiliation, if any, is unknown, because no definitive information was found. There were only five instances of “unknown” affiliations in our study.

The affiliation information was supplemented with the Carnegie Basic classification numbers[15] of the institutions where applicable, an indication as to whether the institution is foreign or domestic, whether the institution is home to an AAUP-member press, and whether the institution is a liberal arts college.[16]

The authors of 364 of the 2579 publications were affiliated with foreign institutions that are not AAUP members. (Another seventy-five are affiliated with foreign AAUP members, and hence are not encoded as “free riders”). Forty-three publications were by museums, government agencies, and other non-academic domestic entities.

AAUP Publication of Works from Non-AAUP Sources (“Free Riding”)

Of the 2,579 publications that we examined, 1,161, or 45 percent, were authored by faculty who are affiliated with AAUP institutions, foreign and domestic, implying that 55 percent of publications were authored by “free riders.” (We are assuming that the five authors that we could not find are not faculty at AAUP institutions.) There are any number of details of our study and the data on which it is based that would lead to changes in this estimate, but it is unlikely that our summary estimate of the prevalence of “free-riding” as defined here would change very much, although the fact that we have oversampled Group 4, which has the lowest incidence of “free-riders,” suggests that if anything, our estimate of “free riding” may be somewhat low.

Going into detail by AAUP Group and Carnegie classification, we see that the aggregate estimate masks a good deal of interesting variation (see Appendices for details):

The incidence of “free riding” by AAUP Group ranges from a high of 74 percent for Group 1 to a low of 46 percent for Group 4, with Group 2 at 62 percent and Group 3 at 55 percent. Because Group 3 and Group 4 account for the lion’s share of all publication, the weighted average comes to 55 percent as noted above.

Looking at the data by Carnegie classification, we find that the great majority (80 percent) of publications from class 15 (Very High Research Activity) are authored by faculty who are affiliated with AAUP institutions. The share falls to 52 percent for the High Research Activity group (class 16) and is nowhere near as high as 50 percent for any other Carnegie class. (Note that the rates of free riding in the following chart are the complements of the shares of faculty who are affiliated with AAUP institutions. Thus, e.g., 80 percent of publications from class 15 are from AAUP institutions, and 20 percent are free riders.)

Although Carnegie Classification 15 (what we used to call Research I) has the lowest incidence of free-riding of any set of domestic academic institutions, it accounts for a largest fraction of free rides from such institutions (16 percent), because it accounts for such a large fraction of publications in total. Large master’s level institutions (Carnegie 18) and liberal arts baccalaureate institutions (Carnegie 21) account for, respectively, 12 percent and 11 percent of all “free riders.” The biggest share of “free rides” comes from foreign institutions (26 percent), followed by the group of authors for whom there is no relevant institutional affiliation (18 percent). Between them, these two groups of authors, with no direct connection to US colleges and universities, account for nearly half of all “free rides.”

Concluding Notes

There are many ways that we could improve this study. We have already explained that we would prefer to do a random sample of publications rather than a stratified sample of AAUP presses. And, no doubt, we could find affiliations for at least some of the five for which we have not found definitive information. Even so, the basic patterns in the data are likely to survive further cleaning and reinterpretation:

  1. Approximately 50 percent of works in this study (55 percent under our assumptions) are currently “free riders” in that their authors are either unaffiliated or affiliated with institutions that are not members of AAUP. Recognizing that most AAUP institutions are subsidized by their home universities, this implies that any broad-based system of publication subvention based on author affiliation could lead to a consequential change in the extent to which support for research is aligned with support for associated publication.
  2. Works by authors from abroad who are not affiliated with institutions that are members of AAUP constitute about 14 percent of all AAUP publications, and are currently “free riders,” accounting for 26 percent of “free rides.” It seems likely that in an environment of widespread institutional subvention of the kind we have been studying, many of these foreign institutions would choose to participate; however, we have no idea how many, and the numbers are big enough to matter, implying that the question should be studied. A reconfiguration of support for US-based monograph publishing could have a substantial impact on scholars who are based abroad.
  3. Large master’s institutions and liberal arts colleges also have a good deal at stake. Currently they account, respectively, for 12 and 11 percent of all “free rides.” We expect that some of these institutions would want to participate in a direct subvention program in order to promulgate the research of their faculties. They would lose the financial benefit of their “free rides.”
  4. In our study of faculty and institutional considerations involved in developing a system of direct subventions for humanities publications, we noted that many faculty members were concerned that the system would disadvantage a significant population of authors and associated institutions, as well as unaffiliated authors, who currently do not support presses but do publish with AAUP presses. That is, these faculty saw “free riders” as needing support and unlikely to obtain it in a world where direct, institutionally-based subvention is prevalent. At the same time, some faculty and administrators saw reduction in “free riding” as an unambiguously positive feature, and a direct subvention system as a sustainable way to support academic publication undertaken by college and university faculty.

Both of these points of view have merit, and both are articulated at greater length in our earlier report, A Study of Direct Author Subvention for Publishing Humanities Books at Two Universities.[17] What we have shown here is that the magnitude of “free riding” is large, implying that changes in the system that would substantially reduce that magnitude could have material effects on publishers, authors, and authors’ institutions.


We are grateful to Joseph Esposito, Meredith Kahn, and Charles Watkinson for helpful comments, and especially to Joseph Esposito and Karen Barch for making data available to us. We are also grateful to the Andrew W. Mellon Foundation for their support of this project, and to Austin Slaughter and Shelby Stuart for research assistance.

Paul N. Courant is Arthur F. Thurnau Professor, Harold T. Shapiro Collegiate Professor of Public Policy, Professor of Economics, Professor of Information, and Faculty Associate in the Institute for Social Research at the University of Michigan. From 2007-2013 he was University Librarian and Dean of Libraries at the University of Michigan. From 2002-2005 he served as Provost and Executive Vice-President for Academic Affairs, the chief academic officer and the chief budget officer of the University. Courant has authored half a dozen books and over seventy papers covering a broad range of topics in economics and public policy. Recently he has been studying the economics of universities, the economics of libraries and archives, and the changes in the system of scholarly communication that derive from new information technologies.

Terri Geitgey is a Senior Associate Librarian who manages print on demand services for Michigan Publishing Services, a unit within Michigan Publishing at the University of Michigan Library.

Appendix 1 Data Tables

Appendix 2 Carnegie Classifications Table

The Carnegie Classification of Institutions of Higher EducationTM 2015 Update Public File, available at


    1. 1 “AAU-ARL Prospectus for an Institutionally Funded First Book Subvention,” June 12, 2014, aau-arl-prospectus-for-institutionally-funded-first-book-subvention-June2014.pdf available at to text

    2. Ibid., 1.return to text

    3. Raym Crow, “A Rational System for Funding Scholarly Monographs: A white paper prepared for the AAU-ARL Task Force on Scholarly Communications,” Flint Hill, VA: Chain Bridge Group, 2014, aau-arl-white-paper-rational-system-for-funding-scholarly-monographs-2012.pdf available at to text

    4. Jennifer Howard, “Who Ought to Underwrite Publishing Scholars’ Books?”, The Chronicle of Higher Education Wired Campus, June 27, 2014, to text

    5. James Hilton, Carolyn Walters, Paul Courant, Sidonie Smith, Meredith Kahn, Charles Watkinson, Jason Jackson, et al., “A Study of Direct Author Subvention for Publishing Humanities Books at Two Universities: A Report to the Andrew W. Mellon Foundation by Indiana University and University of Michigan,” Sept 15, 2015,, 2return to text

    6. The term “free rider” sometimes connotes impropriety on the part of an entity that receives something of value without paying for it. The phrase is commonly used in reference to the case under discussion here, but we do not mean to imply any impropriety. It is in the nature of scholarly communication as currently configured that subsidies for university presses that are provided by presses’ home institutions are, in part, subsidies to the institutions that do not have presses but whose academic personnel publish with subsidized university presses. In this context, “free riding” is a natural and rational behavior. We put quotation marks around the term “free rider” to remind the reader that we would prefer a term that was equally descriptive but that did not have any connotation of impropriety.return to text

    7. In an email exchange commenting on an earlier draft of this paper, Barbara Kline Pope, president of the Association of American University Presses, suggested that if a press publishes a book by someone who is not part of an institution that is subsidizing another press, there is no “free rider” problem. She asks, “Shouldn't we only be concerned with presses that are subsidized rather than all university presses? If an author gets published by Princeton (no subsidy), and she is from the University of Maryland (with no press), do we care? There is no free rider problem in this case.”  In our view, there is still a “free rider” problem. When someone from Maryland (or from nowhere or from France) publishes with Princeton University Press, they are getting subsidized by Princeton. To be sure, the subsidy is arguably from the Press, which has an endowment, but there is still almost certainly a subsidy. Meanwhile, when someone on the Princeton faculty publishes with Stanford, there is a subsidy from Stanford. Both Stanford and Princeton are contributing to a system of scholarly publishing that generally doesn’t break even on sales. Maryland is free-riding, whether it publishes with Princeton or Yale (which have endowments) or Minnesota (which profits from the Minnesota Multiphasic Personality Inventory) or Michigan, which gets help from the Provost. 

      Email exchanges among Barbara Kline Pope, Paul Courant, and Charles Watkinson, February 25, 2016.return to text

    8. Email from Joseph Esposito to Paul Courant and Terri Geitgey, March 14, 2016.return to text

    9. Hilton et al., “A Study of Direct Author Subvention for Publishing Humanities Books at Two Universities,” 16.return to text

    10. Ibid., 16.return to text

    11. Joseph J. Esposito and Karen Barch, “Monograph Output of American University Presses,” submitted to the Andrew W. Mellon Foundation September 23, 2014.return to text

    12. Email from Joseph Esposito to Paul Courant and Terri Geitgey, March 14, 2016.return to text

    13. AAUP Group 1—annual sales up to $1,500,000; Group 2—annual sales from $1,500,000 to $3,000,000; Group 3—annual sales from $3,000,000 to $6,000,000; Group 4—annual sales over $6,000,000. Kimberly Schmelzinger, “Association of American University Presses Annual Statistics 2012 through 2015” (February 2016), 5.return to text

    14. The biographical information reported on publishers’ web pages may or may not reflect an author’s affiliation at the time the book was published. Without further extensive research, it is therefore impossible to say whether the affiliation listed reflects an author’s current location, the author’s location when the book was published, or something in between. Absent radical change in the joint characteristics of authors and publishers, our use of information from current web pages for data collected over the past five years should have little effect on our findings overall.return to text

    15. See Appendix 2 table.return to text

    16. Liberal arts status was determined using information from the Oberlin Group (, the Consortium of Liberal Arts Colleges (CLAC) (, Council of Public Liberal Arts Colleges (COPLAC) (, Wikipedia (, and self-identification via institutional websites.return to text

    17. James Hilton, Carolyn Walters, Paul Courant, Sidonie Smith, Meredith Kahn, Charles Watkinson, Jason Jackson, et al., “A Study of Direct Author Subvention for Publishing Humanities Books at Two Universities: A Report to the Andrew W. Mellon Foundation by Indiana University and University of Michigan,” Sept 15, 2015, to text