The Problem with Departmental Revenue/Cost (non)Analysis

Originally published June 2017.
All across the country struggling colleges (and universities) are hiring one of several academic consulting firms to help them get a handle on their finances. The ACTUAL problem the institutions face is low enrollment but this is experienced as “this place costs too much to run” (because tuition revenue is below expenses) and like management everywhere their brains turn to cutting labor costs. In the absence of a vision for what the academic program should look like (or in the presence of an unwillingness to put such a vision on the table), they turn to consultants to help them identify where to cut academic programs. One element informing decisions about academic restructuring in general and instructional personnel in particular is the so-called program cost structure analysis.

The basic logic of this analysis is to identify all the faculty FTE that staffs courses in a given area, identify the compensation of these individuals and then add in the cost of the program’s share of administrative support and operating budget and then compare this with the “revenue” the unit generates through crediting a fraction of effective per student tuition for each credit hour earned by students in the program’s courses. Then we either subtract one from the other or form a ratio and characterize the program as “in the black” or “in the red” a “net revenue generator” or a “net cost center,” etc.

Distortion and Bias on the Cost Side

When such analyses use actual faculty salaries rather than average faculty salaries they bias the result by faculty seniority.  Since faculty are sometimes on leave and since senior faculty retire and get replaced by new assistant professors this introduces big year to year distortions that make comparisons problematic.

Suppose biology, for example, has had three senior retirements in recent years all of whom have been replaced by new junior faculty. If we look at the department 4 years back it looks very expensive, if we look at it today it looks very inexpensive.

Now, some will answer this observation saying you have to budget for the actuality of today. Point taken. But the stated purpose of this analysis is to understand the relationship between cost and demand.  We are trying to understand something about the liberal arts college of today. If we do the analysis and find that philosophy is more expensive per student than marketing but the reason is marketing is a brand new department that only just hired faculty last year and we make strategic long term decisions on the basis of this information we are going to be making mistakes.

The solution is simple: use weighted average cost that takes into account the actual distribution of the college faculty across pay levels.  This permits program to program comparisons unbiased on the cost-side of the equation.

Distortion and Bias on the Revenue Side

One piece of the demand and revenue side of the analysis is simply looking at student course registrations – how many students do we teach.

This is a fair measure and it’s not hard to zero in on how many students each faculty member has to teach each year to “pay their salary.” When I did this computation a few years ago it came in at around 95 per year.

But using aggregate course registrations as a measure of student interest is problematic.  Many courses in the curriculum have numerous prerequisites and many courses are mandated as part of various minors and majors and general education schemes. And some courses are scheduled in a manner that reduces the number of potential enrollees (not necessarily out of poor scheduling strategy: languages may need to meet 4 times per week, some courses have required labs and labs may need to take up an entire afternoon).

A course that has an absolute prerequisite will almost necessarily never have more students in it than the prerequisite. Departments and programs that are more hierarchical will offer more courses that are necessarily smaller.  Courses with no prerequisites have a natural advantage. English, for example, has dozens of courses with no prerequisites or only English 1 as a prerequisite, a course that every student is required to take.  This gives the English program a huge advantage over, say, biology or biochemistry.

Programs that manage to control general education requirements and get more of their courses to count for GE will have enrollment numbers inflated over “actual student interest.”

The Upshot

The bottom line is that there are a number of structural distortions that make credit hours generated an invalid measure of student interest, especially in comparisons among close cases.

When both the numerator and denominator in a metric are subject to biases moving in different directions the metric is not a valid measurement of what you think it is a measure of. Employing such a metric for comparisons between programs, development of curricular strategy, and ending instructors’ careers is, at best, problematic.

Sometimes an analysis has a data problem (“garbage in, garbage out”) and that’s probably true here. But the far more serious problem lies in the methodology.

How to Fix

There really is no excuse for not using average faculty compensation, unless we do not care about chopping out a part of the curriculum simply because of when we hired the faculty who teach it. The other problem is much hairier.  The very nature of knowledge affects the results here, as do contemporary ideas about assessment that encourage a pedagogical trajectory from “introduction” through “practice” to “mastery.” Taking into account how different programs manifest these is not easy.  But failing to take them into account undercuts the believability of one’s results.