Editor’s Note: This is a guest post by Kent Anderson, founder of the “Scholarly Kitchen,” who is now publishing a subscription e-newsletter called, “The Geyser.” A version of this post was originally published on April 30, 2019. It has been adapted and updated.
The subscription model has been controversial ever since the Internet went mainstream in the late-1990s. Visions of free information permeated the culture. In scholarly publishing, so-called “Big Deals” have also been a source of friction. Various claims about the burden and effects of subscription expenditures have been made, with most claiming the costs are too high and ultimately unbearable. Some recent high-profile contract negotiations have been predicated on this idea, as have various policy proposals.
In March, I analyzed a set of subscription expenditures for about 900 public and private universities and colleges, working from data published in the Chronicle of Higher Education. My math showed that these academic libraries are devoting about 1/3 of their budgets to subscription purchases, with average cost per title of $23.48 for public universities and colleges, and $18.74 for private universities and colleges. Pooling the entire set, the average cost per title was about $21. My conclusion was:
Given the broader trends — emergence of China as a research powerhouse and absorbing the papers coming from there; organic growth in the US and EU STEM workforces; growth among higher education institutions; and general desirability around publication by more people — these cost data seem to suggest the subscription model is pretty efficient for institutions.
Some would disagree. The burden of the subscription model is one of the main assumptions behind the open access (OA) movement, after all. Now and again, subscription costs at academic libraries have also been blamed for the rising cost of tuition, fees, and attending university in general, despite evidence that overall spending on library budgets as a percentage of university budgets has been falling since the 1980s.
It’s true that the costs of attending college have risen dramatically over the past few decades, but there are many reasons for this — lower levels of government funding to offset tuition; greater demands from students and parents for campus amenities and services; more competition on multiple fronts leading to greater marketing and facilities costs; and, more administrative employees with good salaries and benefits to support all of the above.
Materials expenditures and salaries have been the main drivers of increases in library budgets. “Materials” is an imprecise word, however, and since I have yet to see an analysis of subscription costs as a percentage of university budgets, I decided to try to understand something about this aspect of the “materials budget.”
I supplemented the same data set from the Chronicle of Higher Education with the annual budgets for the 10 institutions spending the most on library materials in each list (private and public universities and colleges), the 10 spending the least in each, and the 10 spending closest to average per set (identifying the institution closest to the average expenditure level, and taking that school and those around it). This approach generated a sample of 60 institutions from the set of about 900.
What did I find?
For the sampling of public institutions, subscription costs amount to 0.41% of total institutional budget. For the sampling of private institutions, subscription costs amount to 0.59% of institutional expenses. Averaged, the two amount to 0.5% of the institutional budgets of these 60 institutions.
The institutions included by the approach described above ranged from Ivy League universities to state universities to private liberal arts colleges to medical colleges to technical colleges. It was a decent sampling of the space in that respect.
There were some highs and lows in the samples, of course.
For public institutions, the highest share of university budget applied to subscriptions was 1.40%. Among the sampled private colleges and universities, the highest share of overall budget appropriated to subscriptions was 1.56%. The lowest share of university expenditures appropriated to subscriptions for public and private institutions was 0.19% and 0.09%, respectively.
Here’s a table of the private and public institutions spending the most on subscriptions, and the resulting percentage this represents against overall budget:
To put these numbers in a more mundane context, at the University of Michigan at Ann Arbor it would cost each student paying advertised in-state undergraduate tuition ($13,856) only $26 per year to have access to all the subscription materials that institution carries. That means $13,830, or 99.81% of their tuition, goes elsewhere.
It’s worth emphasizing who uses all this information. Students. Faculty. Researchers. Scholars. Information and knowledge are core functions of any university, and 0.5% of the budget spent receiving the most cutting-edge and reliable research and scholarly publications seems like a pretty good use of funds.
This analysis appears to confirm one of the biggest benefits of the subscription or “reader pays” model — costs are spread across more participants in the market, making information or services more generally affordable. Models like author-pays Gold OA, which rely on payments from a smaller set of producers, centralize costs and profits (necessary) to a greater degree, increasing budgetary pressures for fewer market participants. Such models are far less scalable and more fragile.
In looking up the budgets for these 60 institutions, I was repeatedly surprised at how big the annual budgets are for these colleges and universities — from tens of millions of dollars for the smallest schools to hundreds of millions for the middle set to billions of dollars for those buying the most subscriptions. I wonder how many of us appreciate how university budgets have expanded over the past few decades. This perception gap is worth pondering further.
There are a number of limitations to this analysis. First, it relies upon sampling. I attempted to offset this limitation by taking three samples from each set, one representing the top, another the middle, and one for the bottom. Looking across these six sets, there wasn’t much difference once the percentages were calculated. Second, finding budgetary information for institutions introduced some inconsistencies of time periods, mostly in the sense that there are different fiscal years, different approaches to updating public sites, and so forth. So, in some cases the budget data is from 2017, while most is 2018 and some is 2019. Note that the Chronicle of Higher Education data was from 2016-2017. Third, nothing was adjusted for inflation, but this effect should be minimal as inflation has been low. Fourth, the budget numbers I entered were rounded to the closest million, which many were anyhow in budget documents, but this caused some minor (but to me, acceptable) imprecisions in some calculated percentages. After all, the message here is about the overall and general trend. If an institution wants a precise number for itself, it’s easy math. Fifth, it is limited to US institutions.
None of this is to say that library budgets aren’t under stress. However, it doesn’t look like subscription costs are stressing institutions themselves, and this discrepancy is worth pondering.
If subscription costs were a main driver of tuition increases and general affordability issues, you’d expect them to represent a major percentage of university expenditures. Based on this sampling analysis, they are not.
So why are libraries under stress if subscriptions aren’t the problem? I think that part of it is because the incredible growth in outputs over the past 20 years due to the entrance of China into the global scientific publishing economy, as well as organic growth in traditional markets, hasn’t been adequately appreciated by the powers that be. This stress is shared by libraries and publishers. According to the most recent “STM Report,” outputs have grown at 4% per year while publisher revenues have only grown at 2%. The rapid increase in outputs has been shown to be a main driver of cost increases for libraries and publishers.
So why aren’t libraries supported adequately in this environment? As noted above, perceptions can become ossified. They certainly don’t self-update. There may be a perception that the world of the library is still the world administrators recall from their own university days. In addition, I’ll point to something Phil Davis wrote in 2016 analyzing ARL data around sources of expense in their libraries:
The problems affecting libraries and professors are the same problems affecting many other institutions that depend upon highly trained professionals — they cost a lot to recruit, train, and retain. And once in the organization, we like to reward these professionals based on experience and seniority. Unfortunately, these individuals don’t get any more productive as they age — a professor in 2016 is no more productive than a professor in 1966 — meaning that they get much more expensive over time. Unless we believe that we should index professors and librarians salaries to the CPI, we’re going to have to accept that professional costs are going to be difficult, if not impossible, to contain.
Davis also suggested a reason why it’s easier to scapegoat publishers than to look at other, closer, and more uncomfortable contributors to the trends:
Library materials represent a tiny percent of overall institutional expenditures. An outsider may be surprised at how much fuss is made over such a small slice of the institution’s pie. But libraries are still perceived as a core function of the institution, standing as beloved architectural statements of their historic value. Perhaps this is why no one wishes to take responsibility, at least in part, for their financial neglect.
Library expenses are complex. Publishers and libraries are both dealing with a volume of papers that has grown with the rise of China, with increased funding, with greater publish-or-perish pressures across more fields, and with the general growth of the scientific disciplines.
It’s worthwhile to pull back and look at the big picture. Rather than blaming publishers for budget woes as they themselves deal with unprecedented volumes of submissions, higher technology and wage costs, and greater marketing and infrastructure demands, these data to me hint that we might want to celebrate the fact that such vast amounts of knowledge are so affordable. From these data, it looks like site licenses are raindrops in the rivers of institutional spending.
It’s actually pretty amazing that thousands of students at colleges and universities — as well as instructors and researchers — can in most cases get access to much of the world’s developing knowledge for less than a penny on every dollar spent by institutions of higher learning. Subscriptions aren’t a problem. They’re an accomplishment of efficiency and practicality.