I call shenanigans, and yes, I've gone to the source[0] the article cites.
First off, the source doesn't even claim that. They measure total expenses on instruction. The instructors might be getting a shitty pay (most do), but the expenses could be high if the instructor:student ratio is high enough.
Secondly, I highly question their data and methodology.
The only relevant data to support the claim you made is in Figure 14. That figure does not explain how the expenses on "instructors" are tallied.
To start, the universities have purely teaching faculty and research faculty that also teach. The latter can be highly salaried, but also bring a lot of money to the university in the form of research grants, which actually pay the high salaries.
Thus research spending and instructional spending are likely mixed in Figure 14.
(It is a well-known fact that most of instruction is down by adjunct faculty at slave wages).
Going further, I don't see a distinction between undergraduate and graduate instruction. A graduate seminar, say, on tropical varieties for non-archimedean analytic spaces that exactly 5 students are enrolled in because it's something they need for their research can be formally listed as a class, but this 1:5 instructor-to-student ratio is not something a typical undergraduate student will ever see. Furthermore, whether the TA's (grad students) are counted in such ratios is unclear.
Now, let's look at the data[1].
From 1970 to 2009, the instructors' salaries have seen a modest increase (about $4K/year) after adjusting for inflation[4], less than 6%.
However, the cost of attendance has more than doubled, inflation adjusted[2].
For the claim in Figure 14 to hold any water, we should expect to see the student:instructor ratio drop by a factor of more than 2 since the 70's.
This simply did not happen[3]. The FTE student-instructor ratio, on the average, stayed at about 16-17.
Therefore, the Figure 14 that the book[0] includes without references to sources, can't be right.
The data I cited suggests that the tuition costs would be half of what they are today if they were tied to instructors' salaries. However, the costs have more than doubled.
(Everyone who would like to dispute this is welcome to dig into the statistics).
[0] https://www.mercatus.org/system/files/helland-tabarrok_why-a...
[1]https://nces.ed.gov/programs/digest/d10/tables/dt10_267.asp
[2]https://nces.ed.gov/programs/digest/d07/tables/dt07_320.asp
[3] https://nces.ed.gov/pubs2012/2012001.pdf - page 377
[4] Inflation calculator: https://www.officialdata.org/1976-dollars-in-2006?amount=924