Indeed. Many large US universities are more accurately labeled as research centers with schools attached.
Because those grants are extremely restricted in what they can pay for, it's not quite accurate to include them in anything like an "available operating revenue" number.
[0]https://resources.finance.duke.edu/resources/docs/Financial_...
https://www.forbes.com/sites/georgecalhoun/2025/02/26/univer...
In the supplementary table cited as the source, 72.1% of the total costs were awarded as direct costs and 27.9% as indirect costs (NIH, around 2020). That means an average overhead rate of 38.7% over all grants. Because some grants (such as equipment grants) have lower overhead rates, the average over grants with a normal overhead rate is higher.
At my (unexceptional public R1) university, the latest negotiated rates are 56.5% for on-campus research and 26% for off-campus research. The latter is lower, because many expenses that are normally covered by indirect costs become direct costs in off-campus research.
There are at least 3 ways I'm aware of to calculate overhead rates, so I suspected people were comparing oranges to apples in some places.
As someone who interacts with R1 research institutions as an adjunct and a prime contractor hiring professors as subs, they are far from efficient (like any large organization). My issue is that the people who are producing the actual value are paid and treated pretty poorly generally, and treated as cattle specifically by people whose contributions to anything of substance are extremely unclear but are quite well paid.
I'm doubtful that the organizations at the top end of the list are 4x more bureaucratically bloated than those at the bottom end of it.
I'm highly confident that much more sophisticated research has much higher indirect cost, because a defining characteristic of "sophisticated research" is that it entails exquisite facilities and equipment that cannot possibly be paid for under individual studies.
Another thing I'm confident of is that Harvard et al have much more talented negotiators than the smaller schools, and I'm sure that plays a role. I would be surprised if it explains the bulk of the discrepancy.
Sophisticated research into particle physics, material science, and (for the last several years) AI does come with significant overhead costs for opex. Sophisticated research into most of computer science, mathematics, and other theoretical scientific disciplines does not, let alone humanities research.
I think a large portion of the difference in overhead rates is due to the last item you are confident of in your list (i.e. Duke can tell NIH "if you want our world class researchers to work on this problem, here are our rates", and some random school cannot).
What doesn't make sense to me is to apply the same overhead burden to a theoretical CS or math research effort where someone basically just needs commodity IT resources and an office and one that requires significant time using a particle accelerator.
> Duke has a F&A rate of 61.5% with the NIH, which means that for every dollar provided to a Duke faculty member conducting research, an additional 61.5 cents is given to the University to compensate for its F&A costs.
This is not an uncommon overhead rate for a large university, and is competitive with overhead rates at the largest government contractors. That doesn't mean it's entirely reasonable or a sign of an efficient operation.
According to https://publicaffairs.vpcomm.umich.edu/key-issues/compensati... (just an example of a public university), it's $376K to executives, $481K to deans, and $152.7K to faculty in FY2013. Deans usually count as ~50% admin, so we could call that $376K + $240.5K = $616.5K to admin and $240.5K + $152.7K = $393.2K to faculty, roughly a 3:2 ratio.
Pay him his professor salary, and he'd never have stepped up to the role.
"All complex systems operate in failure mode 100% of the time." What this means is that systems operate with some of their automatic controls bypassed, and with those processes being carried out manually. The Gimli Glider took off with two broken fuel gauges.
My thought about bureaucracy is that you can automate complex human processes only to a certain point, and then the system needs some manual override capability, and possibly human interfaces, to work. This is what bureaucrats do. The reason why its seems chaotic and inefficient is that the easy stuff has been automated away, leaving only the hard stuff.
I can't vouch for every bureaucratic process, and bureaucrat, being optimally efficient or necessary. But in the past few months, I've observed the hard lesson of what happens when you think you can deal with bureaucracies that you think are wasteful by taking a chainsaw to them. I don't believe in that approach any more, even for dealing with systems that I hate.
I'm also an academic. To me, the primary role of a dean is to insulate me as much as possible from upper admin. I've had deans who are good at this job, and those who either aren't good at it, or think that their job is something else. The ones who are good at what I think their job is ... I'm not sure I'd want to see them get 3–4x my pay, but I'm definitely willing to pay a premium to have someone else deal with upper admin.
In contrast, most undergraduate teaching is done by "adjuncts" for whom the job is essentially gig work. Moreover, professors are considered "faculty" and adjuncts "staff," making it confusing to figure out how many employees of a university are engaged in teaching versus doing other things. For instance a faculty-to-staff ratio would be misleading.
Disclosure: I was an "adjunct" many years ago.
Adjuncts are basically contract labor who in general produce much more revenue than they cost.
>The number of staff and non-tenure track faculty has ballooned dramatically since I arrived at Rice in 2004. I agree with you, that from what I've read elsewhere, it's a common phenomenon at well-resourced institutions.
What I've seen at a number of universities are opportunities to get hired on to things like full-time maintenance staff with better pay, job security, and work-life-balance compared to actual PhDs.
And maybe more likely to be a decades-long career at the same institution, compared to recognized scholars.
I couldn't help but notice this about ten years ago, and UH looks like it is on track too.