Unfortunately, that is not the case for many programs. Many programs rely on grad students for TAships, and often the actual teaching, and in many programs, a substantial amount of the productivity of the department is the work being done by PhD students. Much of the rankings of a department are dependant on grad students, from the number of international students, productivity, as I wrote, citations per faculty, and more.
I do know faculty at some universities who are required to have graduate students, and, in many departments, promotion, and even tenure, are dependant on the faculty member graduating a certain number of PhD students.
Much of academia is built on the fallacy that there should be a constant flow of PhDs into the system, and that the quality of a faculty member as a faculty member is dependant on the number of PhD students they have and that they graduate.
So we have a situation like the Tragedy of the Commons. The benefits are personal, but the costs are shared. Actually, it’s worse, the costs are not evenly shared by academia, but borne mostly by the people who gain no benefit from grad students - people with PhDs who are looking for TT positions.
On the other hand, I would guess that a university like Michigan already has a very strong reputation, and they are not trying to push their rankings up. So they can allow themselves to take a break. Harvard did that with accepting PhD students to its English department. Their faculty are also well regarded, so the impact of this decision on them will be less.
To change the subject, the article cited above does not tell us is how many PhDs are being produced in these universities. While 80% of the faculty are trained in 20% of the doctoral universities, how is the distribution of the number of PhDs produced?
Comparing to the top 50 doctorate-granting institutions from the 2021 Survey of Earned Doctorates, you can see that, while OSU is #7 by number of PhDs granted in 2021, it is only #15 of the top 100 universities for providing faculty members. Worse is U Florida, which is #8 in number of PhDs that are granted, but only #26 at supplying faculty. On the other hand, U Chicago is at #28 for number of PhDs granted, but is #12 among universities from which faculty member come.
This means that the PhDs who cannot find faculty positions are not evenly divided among universities.
Of course, these comparisons cannot be done with places like MIT which has a larger percent of PhD who are in engineering a field in which the majority of PhD do not go to academia. It also misses Caltech, which produces to few PhDs to be on either list, but likely has one of the highest hiring rates of PhDs to faculty positions.
But still, it is very clear that the relative number of faculty from universities tells us little on its own. However, the “prestige” ranking in the article. It is also obvious that PhDs from some universities have a far higher likelihood of ending up with a faculty position.
Ironically, the universities whose graduates who seek faculty positions but are least likely to end up in faculty positions, are those which are least likely to stop producing PhDs.
BTW, Michigan not producing PhDs in a field with a glut of PhDs will not make a difference for the people graduating with PhDs from these universities whose graduates have a more difficult time being hired to faculty positions. The places will be filled by applicants from other universities in that top 20%.