An emeritus professor at the University of Missouri, Arthur Jago, in a recent Chronicle of Higher Education article was skeptical of the claim that a majority of academic journal articles are never read. But Professor Jago accepts that many—maybe most—journal articles are rarely cited. And he notes that some academic journals have circulations of well under 1,000, many copies sitting unread on library shelves. I suspect I have written many academic papers where the number of readers is closer to 10 than 100. This blog will be read by far more readers than the typical meticulously researched paper that I wrote in my academic prime.

In an earlier era, say 1960, faculty at most universities were not expected to do a lot of research. At a mid-quality university, the teaching load might be six courses a year—three classes meeting three hours weekly taught each semester. Today, at a similar university, research is promoted much more vigorously, so teaching loads on average have fallen, say to four courses annually. At research-intensive institutions, the load is lower; at schools with very modest research expectations, it is higher.

The cost of lowering teaching loads to support research is enormous. Supposed an academic department at a fairly large state university thinks it needs to teach 72 courses a year. With a six-course annual load, that means it needs 12 professors (actually closer to 14 to allow for faculty with administrative duties or on sabbatical leave). Suppose in pursuit of research renown, the university cuts its load to four courses annually. It now needs minimally 18 professors to cover the courses (probably 20).

In my field of economics, a professor, even a new assistant professor, costs about $100,000 annually with fringe benefits—often more. Moving from a 12- to 18-member department adds at least $600,000 in costs. Let’s suppose that each of the 18 faculty members publish one paper a year in a reputable academic journal, and that the output would have been nothing were it not for the reduced teaching load. With those assumptions, 18 papers are published for the $600,000 in extra costs associated with research—$33,333 per paper. Suppose 100 people read each paper—probably a high figure. It costs at least $333 for the labor to produce something for every reader.

In fields like history, major scholarly output comes in books. Supposed reduced teaching loads lead to a book being produced by professors every six years (that would be a pretty high rate of output—at least five books in a lifetime). The implicit labor costs associated with each book probably approaches $200,000, for perhaps 200 readers—$1,000 per reader. Moreover, the $200,000 in labor costs probably will yield ten or fewer scholarly citations—at least $20,000 per citation.

There are numerous caveats to the calculations above. The numbers and publication rates and costs vary considerably with academic discipline. The calculation is no doubt meaningfully different in some of the hard sciences, and occasionally commercially successful research discoveries yield big dividends.

The point, however, is that lower teaching loads over the past half-century have come at a cost. In part that cost is met by moving away from using expensive tenure-track faculty to teach, going to more part-time adjunct faculty (some would argue with a loss in instructional quality). The emphasis on research, however, is expensive and leads often to a trivial addition to the stock of knowledge. Diminishing returns set into research. The 100th paper on the impact of gerrymandering, interpretation of John Rawls’ veil of ignorance, or the effects of climate change on the mating habits of insects no doubt is typically less consequential than the first or second paper. Moreover, while research sometimes complements teaching and leads to more insightful instruction, a publish or perish environment can lead to student neglect, perfunctory teaching, and a decline in mentoring of young persons that helps them navigate the transition from adolescence to adulthood.

College costs are getting too high. Benefits are falling relative to costs. Higher education is a labor-intensive enterprise, and cost savings have to come by lowering labor expenses. While a huge hunk of that involves curtaining the burgeoning administrative bloat, a more healthy balance of teaching and research responsibilities probably is in order. New teaching technologies possibly can help, but I doubt Western civilization would be imperiled much if the number of academic journal articles fell by 25 or even 50%.