During 2015, researchers from Indiana University and the University of Michigan collaborated on a project supported by the Andrew W. Mellon Foundation to explore the potential impact of moving to a model where the full costs of scholarly monograph publishing would be covered by the parent institution of the author. This paper reports on one part of this larger project that focused on establishing a framework to catalog and understand the costs associated with monograph publication using data from Indiana University Press and University of Michigan Press. Understanding how much scholarly monographs cost to produce is essential to any system-wide business modeling, but there is little published work on this topic that carefully documents methods of arriving at numbers as well as the numbers themselves. In focusing on the press operations of Michigan and Indiana as case studies, this paper hopes to provide an evidence-based approach to identifying and delineating the costs specific to scholarly book publishing, while offering a replicable methodology for discerning with reasonable accuracy the “true” cost of publishing an academic monograph.


Within the Academy, and particularly within the humanities, the scholarly publishing of books is one of the few ways academics can distinguish themselves for tenured positions. Despite the importance of book publishing to the functioning of a robust ecosystem of scholarship, however, the business aspects of publishing scholarly books have been made opaque in a market-based system where publishers act as competitors. It is only recently, as academic book sales have reached what many consider unsustainable levels that the actual costs of book publishing have come under scrutiny.[1] A particular push for transparency has come from organizations concerned about the health of the academic ecosystem and interested in exploring different ways of funding scholarly publishing. These include libraries, government entities, academic institutions, and private foundations.[2]

The main driver in the United States has been the publication of a study by a joint committee of Provosts (from institutions that are members of the American Association of Universities, AAU) and Directors of Libraries (from libraries that are members of the Association of Research Libraries, ARL) on the feasibility of “flipping” the current funding model for scholarly monographs. [3] This prospectus, compiled by consultant Raym Crow, proposes that the cost of publishing first books might be fully funded by the parent institutions of the author(s) rather than through sales and licensing income from consumers, with the final version being made available open access as well as in premium print and electronic editions. Crow uses a notional figure of $20,000 per book to model the potential costs of such a system, but notes that this is a convenient estimate rather than a data based number. A number of issues to be addressed prior to the potential implementation of such a system are explored in the broader IU/UM report from which this project derives and in a similar exploration conducted by Emory University.[4]

So it is in this context of increasing interest in the economics of publishing that this investigation attempts to answer a basic question: What are the full costs of publishing a scholarly monograph at university presses based at Indiana University and the University of Michigan? In financial terms, the lead researchers on this project (a business professor and his graduate student) conducted an “audit” of publishing within these two publishing houses in an effort to better understand the costs associated with publishing a scholarly monograph.

There have been a number of investigations on the costs of publishing scholarly journal articles well reviewed by Donald W. King.[5] However, robust studies into the costs of producing academic books in a North American context have been lacking until recently. A notable exception is the work of Marlie Wasserman, but this is now somewhat dated.[6] A more recent study by Eelco Ferwerda and colleagues at the OAPEN Foundation is useful but inconclusive.[7]

Driven by the AAU/ARL proposal, the situation has changed with two studies working in parallel providing deep dives into questions of cost. The project reported in this paper is one, but the second, also funded by the Mellon Foundation, reports on an impressive research program led by Nancy Maron on behalf of Ithaka S&R.[8] For her project, Maron and her colleague Kimberly Schmelzinger visited a sample of twenty university presses, which were diverse in size and location, and studied not only the direct costs of monograph production but also the indirect costs of time spent by press staff. Maron’s final report paints a detailed and nuanced picture of variations in costs for monographs published in 2014 and discusses the bases for these differences but necessarily obscures the identity of specific presses studied. By digging into the general ledgers of two identified publishers also studied by Maron, drawing on the same 2014 data that she did, but reconstructing costs using a different methodology we hope that this paper can be seen as complementary to the Ithaka S&R study. Some comparison of findings between this report and the work of both Wasserman and Maron is shared in the “Discussion” section below.


A couple of clarifications about the process that we (Smart and Fitzgerald) used to conduct the financial analysis are worth making at the start. The charge of our investigation was not to develop a “profit-and-loss” statement for the press; in other words, revenue and other revenue-related considerations (such as royalties, which are a function of revenue) were identified, considered, and ultimately removed from the final analysis, keeping in mind that our goal was simply to determine, on a per-monograph basis, the “true” cost of publication. Also, as made clear by the broader Mellon grant of which this investigation was part, many universities support the ongoing operations of their presses through internal subsidies, and we did not take these into account when determining the cost of scholarly publishing at the per-monograph level.

Differences in type of publishing

Our first step was to identify the different types of published titles at each press. At Indiana University, for example, there are two primary types of titles published: 1) so-called trade books (e.g., “coffee-table” books on Indiana history, or a photobook of Indiana University’s campus, written primarily for a non-academic audience); and 2) scholarly monographs (works of scholarship written by academics, for academics, for the primary purpose of distinguishing the academic credentials of the author and thus adding prestige to the university’s imprint). Different from scholarly monographs, trade books typically contain detailed visuals, more graphically interesting typography, and various charts and graphs, all of which make these books, in general, more costly to publish than scholarly monographs. (And even within the realm of trade books there existed a range of high-, medium-, and low-cost trade books, whose relative weight within the overall trade output of the press had to be considered.)

At the outset, then, we felt it was important to have an understanding of the mix of books each press published (monograph vs. trade vs. any others), and how (or if) the existing mix differed from what the director of each press desired to produce in any given year.

Data sources

Rather than using each press’s finalized financial statements as a starting point, we chose instead to focus on each press’s general ledger: the statement of all financial transactions, inflows, and outflows. Although we would have ideally liked to use multiple years’ worth of data, changing accounting conventions over the last few years at both presses meant that only data from fiscal year 2013–2014 was easily available (also the base data for the Ithaka S&R study).

Typically, a general ledger comes in the form of an Excel spreadsheet with tens of thousands of individual line-items. We worked closely with the accounting teams at each press to ensure a complete understanding of the data, and, once comfortable, we manually “rebuilt” the income statement of each press by aggregating line items by financial “code,” that is, whatever combination of letters and digits each press uses to catalog and differentiate between entries in the general ledger (e.g., travel expenses are given one code apart from expenses for office supplies). By beginning in this way, we accomplished three important objectives critical to establishing a proper foundation for our investigation: 1) verify the completeness of each press’s reported financial condition as expressed in the finalized financial statements against the “raw data” contained in the general ledger; 2) gain an important understanding of what costs in the aggregate were “material,” that is, both significant in relative size and also critical to the operations of the press; and 3) clearly determine the impact of allocation of material costs on total monograph cost.

A note on materiality (as defined in 2 above): Because we were able to aggregate costs and sort them by size (thus determining the impact of each type of cost on both presses’ overall cost structure), it was fairly easy to identify a proper threshold for cost importance. In the aggregate, costs below six figures were not considered material, because every cost in aggregate above $100,000 comprised approximately 80 percent of total operational cost at each press. These only included three or four cost categories.

A “steady-state” year

Despite only one year of workable data, our best efforts were made to “normalize” the data to reflect at each university press what its directors—Charles Watkinson (Michigan) and Gary Dunham (Indiana)—would define as a “steady-state” or “typical” circumstance at each press. After discussing this concept with both directors, we discovered that in order to accurately recreate a meaningful steady-state model, we would have to focus particular attention on two cost categories: 1) labor; and 2) total press output.

Regarding labor: Both Michigan’s and Indiana’s presses were without a director for the majority of the fiscal year for which we had data. In addition, there were several press employees who were not employed the full year or whose positions were vacant, as well as positions for which the press had not yet hired, but had concrete plans to do so. But despite the lack of “fully-staffed” data, we felt it would be unreasonable to attempt to calculate the cost of publishing a scholarly monograph without including and properly allocating the full-time salary and benefits for individuals, including the director, whose activities in the process of monograph publication are fundamental to the long-term operation (and therefore cost structure) of the press. It is not as if the press could operate in perpetuity with only, say, one acquisitions editor when in reality it needs two—even though it may have done so for a short period of time.

Regarding total press output: Consider that directors of university presses set objectives for their press’s output in a given year. For example, suppose that on December 31, Dr. Dunham sets a goal of publishing 135 books at Indiana during the course of the following year, 90 of which are to be scholarly monographs. Then suppose in that following year Indiana University Press only publishes one hundred books, sixty-two of which are scholarly monographs. It was insufficient in our minds to simply take that reduced output and proceed from there. The real value of this investigation, we hope, will be in providing press directors with the true “all-in” cost of publishing a scholarly monograph—not something less than that. If we effectively ignored in our analysis what a “full-output” cost structure might look like, this fails to account for the differences between fixed and variable costs in each press’s cost structure and the respective impact that each type of cost has on the press’s overall financial performance. It also ignores the desired mix of books for each press and the financial impact of that mix.

Additionally, we believed that if we were comfortable “normalizing” the press’s labor costs under a fully-staffed model, we should do likewise in terms of a full-output model. This requires an understanding of what overhead costs in each press’s structure were in fact fixed vs. variable, as well as what the “average” cost and mix of existing titles were in the year for which we have data. Just as we inserted full-time salary and benefits figures in our analysis for positions that may not have been filled at the time, we would need to understand average figures for overall titles at each press if we were to conduct an analysis as if the press had printed, in the above example, twenty-eight more monographs than it actually had.

Cost allocation

It is important to keep in mind that, as mentioned earlier, both presses produce more than one type of published title, whether journals, monographs, trade books, or others. Given that our investigation involved determining the cost of scholarly monographs only, we needed to create a defensible methodology for separating and allocating costs that may inevitably overlap as part of regular press operations. For example, an acquisitions editor may spend some of his or her time on journals and some on scholarly monographs. The question is, how much time? And, returning again to our effort to “normalize” our results, is that time considered “typical”?

With this in mind, two (material) cost categories stood out as more difficult in terms of our ability to “unwind” them; however, we analyzed these categories to determine what proportions of each could properly be allocated to the publication of scholarly monographs:

  1. Labor costs. Theoretically, at both Michigan’s and Indiana’s presses, there would be individuals who worked 100 percent on monographs, 0 percent on monographs, and everywhere in between. We decided the best way to approach this was to interview the directors of each press and ask them what the ideal mix of each employee’s time should be on monographs vs. other types of publications. Following this, we worked closely with the accounting and human resources teams at each press to collect salary and benefits data for each employee and, based on the breakdown given to us by the press directors, separated each employee’s total compensation into two buckets: “monograph” and “other.” For example, if an employee at Indiana University Press made $100,000 in total salary and benefits and, ideally, 75 percent of her time should be spent on scholarly monographs in a given year, we allocated $75,000 of that employee’s labor cost to monographs. This process was repeated for every professional employee at each press.
  2. Physical printing costs. Printing posed a similar problem, not only because each press publishes a variety of publications but also because each publication type has a different cost structure (this was touched on briefly at the beginning of the “Process” section). Using Indiana University as an example, trade books comprise a significantly large portion of cost at that press. And within trade, we were told that there are high-cost, medium-cost, and low-cost trade books. We were given an average number of each cost-type of trade book that Indiana’s press would, under ideal circumstances, publish in a year, and thus derived “weights” for each cost-type of book in the press’s trade portfolio. For example, of the 45 books Indiana University’s press published in fiscal year 2014, 15 (33 percent) were deemed high-cost, 10 (22 percent) medium-cost, and 20 (44 percent) low-cost. From there, we took a random sample of each cost-type of trade book (high, medium, and low) to determine an average individual print cost, and then applied these average costs to the weights we derived for each cost-type. In this way, we were able to determine an “average total cost” for desired trade printing at Indiana in a given year.

For scholarly monographs, we took a random sample of thirty recently published titles and averaged their printing costs. We then took that average per-monograph printing cost and applied it to the number of monographs that Indiana would, under ideal circumstances, publish in a year. In this way, we were able to determine an “average total cost” for scholarly monograph printing at Indiana in a given year.

By summing these average total costs across publication type (trade and monograph), we could generally determine the weight of monograph printing cost in the overall mix of the press’s total printing costs. This number was 60 percent. (Initially, we had assumed that because Indiana publishes generally two-thirds monographs and one-third trade books, that number would be closer to 67 percent.) We applied a similar approach to Michigan’s press where appropriate and where applicable. For all other costs at both presses (those considered immaterial being under $100,000 in the aggregate or those for which we did not have sufficiently detailed data) we applied a monograph/other split based on the desired overall mix of publications at each press (90 percent monograph / 10 percent other at Michigan; 67 percent monograph / 33 percent other at Indiana).


Once all costs had been properly allocated, and we knew the total number of scholarly monographs desired at each press, we arrived at an initial per-monograph cost figure:

University of Michigan – Ann ArborIndiana University

We found this disparity troublesome; it seemed that the activity of the presses were reasonably similar, so a 23% difference in cost seemed inaccurate. We thought at first that it may be because Indiana employs about 25 percent more people than Michigan, but each presses’ total labor cost (compensation and benefits) was nearly identical.

After further discussion with Dr. Dunham at Indiana, we realized that we had not properly accounted for another type of publication: scholarly journals. Because both monographs and journals are in the “scholarly” category of publication, there was a significant portion of journal cost (in fact two-thirds of it) that had already been accounted for. We were given total journal cost figures from Indiana’s accounting team, and after removing two-thirds of that cost, and then dividing by the total number of monographs, we arrived at a new per-monograph cost figure:

University of Michigan – Ann ArborIndiana University

After sharing the methodology and the figures above with full teams at both the University of Michigan and Indiana University via conference call, we learned that what would be even more useful to university press directors and practitioners would be an understanding of the cost of the “zeroeth” copy of a scholarly monograph: that is, all of the costs associated with transforming an approved manuscript to a published scholarly monograph up until printing. This immediately excluded two substantially material cost categories: printing and royalties.

Because of the work we had already performed on printing cost allocation as explained above, removing the printing costs associated with monographs was an easy task. Regarding royalties, we went back to the accounting teams at both presses and determined a proper royalty allocation rate for scholarly monographs.

Additionally: Another consideration that we did not immediately take into account involved the “rent” that each press pays to the university for the use of its physical space. On a per-monograph basis, this is typically a negligible allocation (approximately $1,000 at Michigan and under $500 at Indiana), though individuals wishing to recreate this type of analysis at their own press should not fail to take allocated cost of space into account.

Once we subtracted printing and royalty costs, and added in the allocated cost of space, we arrived at the following per-monograph cost figures:

University of Michigan – Ann ArborIndiana University

It is encouraging that these figures are close to the studies by Wasserman (1998) and Maron et al. (2016) cited earlier in this paper. Wasserman provided “operating expense” numbers on a per-title basis for two individual monographs. These came in at $22,075 and $23,499. Assuming average inflation of 1 percent per year between 1998 and 2015, these numbers grow to $26,144 and $27,830—nearly identical to ours. Maron et al. reported a mean “basic cost” of $26,200 and a mean “full cost” of $34,098 for Group 3 presses including Indiana and Michigan. Individualized numbers for Michigan were $23,391 for “basic cost” and $26,022 for “full cost” and Indiana’s numbers were similar. Maron’s study used a different and more complex methodology but focused on the same sample of 2014 books and attempted in the “full cost” to elicit a similar picture of direct costs plus overhead, so the near match is encouraging. More information on the methodology of Maron’s study is described in their white paper.


As well as contributing some evidence-based numbers to the conversation about monograph costs, the goal of this paper was to explain and demonstrate a reasonable, replicable methodology for determining publishing costs and to help press directors and others within scholarly publishing to better understand their own press operations.

In the interests of replicability, we offer some recommendations for those who would want to conduct a similar investigation at their own university press:

  1. Be sure to fully understand the landscape of your university press: the steps a manuscript must take before becoming a published monograph; the number and diversity of individuals who work on the manuscript throughout the process; and the types of publications (other than monographs) that your press publishes.
  2. Don’t use the finalized financial statements as a starting point. Dig into your press’s general ledger: it will be invaluable to you later and will get you far more familiar with the operations of the press at a necessary level of granularity. Rebuild the income statement from there, and lean heavily on your press’s accounting team for context, insight, and guidance in doing so.
  3. When allocating costs among the various types of publications at your press, be consistent and methodical throughout. Every decision need not be perfect but must be, at minimum, defensible.
  4. By far, labor is the largest expense at any successfully functioning press. Be sure to talk to employees at the press about how they spend their time among the various types of publications they work on. Also, while you may believe that there is no such thing as a typical year when it comes to employee turnover or title output, it is important to conduct the analysis as if there were.
  5. Understand what’s more important to those to whom you will eventually deliver your analysis: the cost of publishing a scholarly monograph into the future (that is, including printing and royalty costs); the cost of a “zeroeth” copy; or, something else. Be prepared to adjust your analysis accordingly.


The authors would like to thank the Andrew W. Mellon Foundation for supporting the project to conduct “A Study of Direct Author Subvention for Publishing Humanities Books at Two Universities” (PI: Carolyn Walters, grant 41400692 made in December 2014) of which this paper is a product. We would also like to thank the entire staff of the presses of the University of Michigan and Indiana University, especially: Katie O’Brien; Carolyn Walters; Lisa Hartwell; and Gabriela Beres. We could not have done this without you.

Scott Smart is clinical professor of finance at Indiana University in Bloomington, where he is also the associate chair of the full-time MBA program, the director of the Strategic Finance Academy, and the Whirlpool finance faculty fellow. His research focuses primarily on initial public offerings, mergers and acquisitions, and corporate governance, and has been cited by Business Week, the Wall Street Journal, Fortune, the Harvard Law School Forum on Corporate Governance and Financial Regulation, and the New York Post. Scott received his PhD in finance from Stanford University. He can be reached via email at ssmart@indiana.edu.

Charles Watkinson is director of University of Michigan Press and associate university librarian for publishing at the University of Michigan Library. He was previously Director of Purdue University Press and Director of Publications at the American School of Classical Studies. Charles received his MBA from Oxford Brookes University. He can be reached via email at watkinc@umich.edu

Gary Dunham is director of Indiana University Press. He was previously director of publications for the American Speech-Language-Hearing Association, executive director the State University of New York Press, and director of University of Nebraska Press. Gary received his PhD in anthropology from the University of Virginia. He can be reached via email at dunhamg@indiana.edu

Nicholas Fitzgerald is a 2015 MBA graduate of the Kelley School of Business at Indiana University, where he studied finance and accounting. He currently works at Target Corp. at its headquarters in Minneapolis, Minnesota, where he is a lead financial analyst within the company’s Financial Leadership Development Program. He can be reached via email at ncfitzgerald@gmail.com


    1. As John Thompson writes, “In the 1970s academic publishes would commonly print between 2,000 and 3,000 hardback copies of a scholarly monograph . . . Today many academic publishers say that total sales of hardback-only monographs are often as low as 400–500 copies worldwide” Thompson, J. (2005) Books in the Digital Age. Cambridge: Polity Press, pp. 93–94.return to text

    2. A good review of the issues is the report commissioned by the Higher Education Funding Council of England (HEFCE) from Professor Geoffrey Crossick; Crossick, G. (2015) Monographs and Open Access: A Report to HEFCE. (January 2015), http://www.hefce.ac.uk/pubs/rereports/year/2015/monographs/return to text

    3. Crow, R. (2014). AAU/ARL Prospectus for an Institutionally Funded First-book Subvention. (June 2014),http://www.arl.org/storage/documents/publications/aau-arl-prospectus-for-institutionally-funded-first-book-subvention-june2014.pdfreturn to text

    4. Walters, C. et al. (2015). A Study of Direct Author Subvention for Publishing Humanities Books at Two Universities: A Report to the Andrew W. Mellon Foundation by Indiana University & University of Michigan. (Sept 15, 2015), http://hdl.handle.net/2027.42/113671 and Elliott, M. et al. The Future of the Monograph in the Digital Era: A Report to the Andrew W. Mellon Foundation by Emory University. (July 1, 2015), https://pid.emory.edu/ark:/25593/q4fd0 return to text

    5. King, D. W. (2007). “The Cost of Journal Publishing: A Literature Review and Commentary,” Learned Publishing, 20(2), 85–106.return to text

    6. Wasserman, M. (1998). “How Much Does It Cost to Publish a Monograph and Why?” Journal of Electronic Publishing, 4(1).return to text

    7. Ferwerda, E., Snijder, R., & Adema, J. (2013). Oapen-nl: A Project Exploring Open Access Monograph Publishing in the Netherlands: Final Report Oapen Foundation.return to text

    8. Maron, N., Mulhern, C., Rossman, D., and Schmelzinger, K. (2016). The Costs of Publishing Monographs: Toward a Transparent Methodology, Ithaka S&R.return to text