An academic CS department has to attract PhD level talent in hot areas like ML while only paying $30k a year.
On the purely financial side, many top students work internships during the summers and make 40 or 50 thousand dollars during that time in addition to their academic stipend.
Compared to almost any activity a university could take, it is incredibly cheap to bring in mathematics PhD students.
They don't attend classes after ~2 years, mostly operate independently besides consulting with their advisor, don't take anywhere near a professor's salary, teach or TA classes, act as lab technicians, and bring in money through grants.
The costs are mostly upfront in the form of providing the necessary research facilities to attract research-oriented faculty and students, and the administrative staff needed to ensure compliance with grant terms.
In America the students at the undergraduate and masters levels pay to pursue their degrees, while the PhDs are paid by the school. As these students do not directly generate revenue, the PhD programs will be first on the chopping block and the admin who make the 'tough decisions' to keep the ship afloat will be off at their next jobs by the time the chickens come home to roost.
The alternative would be hiring dedicated employees to help with grading and lab sessions, and they won't tolerate the $30k/yr a PhD student does. This would have immediate impacts too, as there's no way a lone professor can keep up with grading for the class sizes in early undergrad.
It sounds like people don't understand bureaucracy is always imperfect. If it was perfect then you don't need to create another agency called DOGE while having Congressional Budget Office and do exactly the same things.
The question should be is there fraud/fat/waste which has a meaningful impact? If not then it changing it wouldn't really matter. The unfortunate thing is that anecdotal evidence rules supreme and there are enemies every where.
"Data doesn't support a meaningful impact? I saw it with my own eyes so it should be true and the person reporting the data must have Democrat agenda"
What does a graduate math program need? A building with some offices and classrooms, wifi, email service, maybe a couple of secretaries and janitors, office supplies, and salaries for students and researchers/instructors.
What need does a math program have for any but the most basic administration? That's where all the money is going, where the biggest growth in spending is going.
You could cut university admin costs by 75% and lose nothing. Start with the top 25 university presidents who all earn a slightly rounded up 2 million a year and more.
Most of the time, when you hear a politician saying that universities should / should not do X, they are effectively saying that universities should spend more on administration.
Universities with a residential campus have a lot of staff in functions unrelated to the academic mission, such as student housing, food services, healthcare, or sports facilities. If they have to compete for students instead of most people just automatically choosing the nearest university, focusing on these tends to make them more competitive. And while student amenities are not particularly important to PhD students, they are important to the university if it also educates undergraduates.
Then there is the organization chart. In a traditional university, the faculty senate (or another similar body) is in charge and all administrators are subordinate to it. But the modern world prefers centralized organizations, with administrators at the top. And whoever is in charge also determines the priorities of the organization.
People say that, but could you really? I'd love to see a breakdown on how you pull this off.
There are extreme downsides - for many colleges athletics is a money-maker. So is administering IP. There's is also lots of real estate, which appreciates value, but needs maintenance.
The big reason for all the extras is that it makes the school known, in a very big and important way. They host conferences, have archives that receive donations, give out awards.
Donors give all sorts of weird tasks - and funding to achieve them.
Modern colleges are so many different things.
There is a subset of colleges that adopt the "keep it simple" approach, but they often run into lots of trouble. The big problem is without doing tons of stuff, people forget they exist.
It's a bit like drug companies advertising at the super bowl. They hate doing it, but don't have a choice.
University admin work expands to the available workforce and I've heard first person accounts from long time senior university staff about admin employees who literally didn't do anything of any conceivable consequence.
