A hypothetical example – A business school is offering a non-degree executive education course.  Forty people sign up and yet the course is cancelled.  An administrator says that when the overhead costs are taken into account, the course is unprofitable.

Without the class, the lecture rooms and dorm rooms would be empty and the same number of staff employed.  That is, the overhead cost is incurred whether the class is taught or not.  Hence it should not be included in determination of the cancellation decision.  The revenue from the class should be compared with the extra variable costs incurred from running it (e.g. the salary paid to the teacher!).

The reverse fallacy is also possible: The class is on the schedule, very few students have signed up and yet it is not cancelled. If it is cancelled, the logic goes, how would the school recover the overhead?

Jeff and I have a paper Mnemonomics… that offers a theory of the sunk cost fallacy based on limited memory.  We call the first fallacy above “the pro-rata fallacy” and the second “the Concorde fallacy” (after the supersonic jet).  In experiments, we find the pro-rata fallacy is extremely common.