"Building new lamps" is why the engineers and technicians developing new scientific instruments and experimental processes deserve more credit than they get.
Yes- after many years of being a theory guy I actually did a 180 and started building my own scientific instruments, because the acquisition cost of a research microscope is so high. This allows me to experiment quickly with new ML algorithms, and I've greatly increased my respect for the people who toil to make the hardware for next-generation discovery science.
59% indirect research costs for administrative overhead seems high. Could it be that these charges against grants are used to fund students in other subject areas where grants are not available?
It's pretty typical, actually. 50% is about the minimum that major universities take out of a grant you get as a researcher at the university.
It's nominally to fund general facilities, etc. At least at public universities, it does wind up indirectly supporting departments that get less grant money or (more commonly) just general overhead/funds. However, it's not explicitly for that. It's just that universities take at least half of any grant you get. There's a reason large research programs are pushed for at both private and public universities. They do bring in a lot of cash that can go to a lot of things.
This also factors a lot into postdocs vs grad students. In addition to the ~50% that the university takes, you then need to pay your grad student's tuition out of the grant. At some universities, that will be the full, out-of-state/unsupported rate. At others, it will be the minimum in-state rate. Then you also pay a grad student's (meager) salary out of the grant. However, for a post-doc, you only pay their (less meager, but still not great) salary. So you get a lot more bang for your buck out of post-docs than grad students, for better or worse. This has led to ~10 years of post-doc positions being pretty typical post PhD in a lot of fields.
With all that said, I know it sounds "greedy", but universities really do provide a lot that it's reasonable to take large portions of grants for. ~50% has always seemed high to me, but I do feel that the institution and facilities really provide value. E.g. things like "oh, hey, my fancy instrument needs a chilled water supply and the university has that in-place", as well as less tangible things like "large concentration of unique skillsets". I'm not sure it justifies 50% grant overhead, but before folks get out their pitchforks, universities really do provide a lot of value for that percentage of grant money they're taking.
Well that makes it sound worse than I thought. Why should it be any higher than the pro rata allocation of the project’s actual use of university facilities (lab space, equipment, etc)?
Even in the defense industry, a cost-plus contract with a 10% margin is a lot. And it’s a federal crime to include costs in the overhead amount that aren’t traceable to the actual project.
In the defense/other industries, everything is put under the "cost" part. There's just a lot more line items that cover all that stuff.
The overhead simplifies this to a large extent in that the PI only needs to account, as a "direct" expense the cost of his team (salaries, and things not covered by the university such as compensation to human subject volunteers, etc.)
Its essentially a subsidy, and been abused for years.
One clarifying point. Indirect is normally charged on top whatever the PI gets. So they don't "take out" 50% the total. They add 50% to the original grant. So if a researcher gets a $500k grant, 50% indirect would be $250k, and the total allocation is $750k.
The indirect is a negotiated flat rate that covers costs that would be too numerous or difficult to account for in the direct costs. Like how would you as a researcher budget a fractionalized portion of access to a supercomputer cluster in each and every grant you need? You would need to hire new accountants just to handle this!
The indirect rate is basically covering the whole infrastructure of research at a university. In theory all could be put into direct costs but…again…we get to tremendously difficult accounting
First the rate was negotiated on a per institution basis with the government. It’s based around a mountain of oversight and compliance. Ironically all that compliance work contributes to the need for more administration.
Second, modern research needs a lot of people doing non-directly research adjacent stuff. Imagine looking at all the support people on an airbase, and saying why don’t we just cut them and let the pilots fly without all this logistics baggage.
If you are Bozo University that has no grants, you also have no overhead, because everything you spend is attributable to that first grant. You spend $50 for tiny little flasks of liquid nitrogen. You buy paper at Staples.
If you are UCSF, you have 80% "overhead" because everything is centralized. Your LN2 is delivered by barge. You buy paper from International Paper, net 20, by the cubic meter. You have a central office that washes all the glassware. Your mouse experiments share veterinarians. All of this costs much, much less because of the "overhead".
If I understand you correctly, what you're claiming is:
University 1 gets $100. $10 of it goes to admin, $90 to researcher. Researcher spends $60 on supplies and equipment. This is accounted for as 10% administrative overhead.
University 2 gets $100. $20 of it goes to admin, $40 to researcher, $40 to supplies and equipment. This is accounted for as 60% administrative overhead.
Is this an accurate characterization of your claim?
> University 2 gets $100. $20 of it goes to admin, $40 to researcher, $40 to supplies and equipment.
This is not how it works; this would be 150% overhead. ($60 / $40).
Basically, if something is a shared utility (common lab maintenance, supplies that can't be metered and charged to specific projects, libraries on campus, etc.) then it's overhead.
Also included in overhead is administrative & HR expenses... and things like institutional review boards, audit and documentation and legal services needed to show compliance with grant conditions.
The reasons for high overhead are threefold:
1. Self-serving administrative bloat at universities and labs. We all agree this is bad.
2. Shared services in complex research institutions (IRBs, equipment maintenance, supplies, facilities). This is good overhead. We want more of this stuff, though we want it to be efficiently spent, too.
3. Excessive requirements and conditions on grants that require a lot of bodies to look at them. This is bad, too, but doesn't get fixed by just lopping down the overhead number.
Unfortunately, if you just take overhead allowance away suddenly, I think it's just #2 which suffers, along with a general decrease in research. Getting rid of #1 and #3 is a more nuanced process requiring us to remove the incentives for administration growth on both the federal and university side.
if a grant is the same $1M and Bozo University gets to spend all million on the actual research at hand, but UCSF only gets 200K, how is UCSF more efficient?
Wouldn't the LN2 be traceable to the project either way as direct non-overhead cost, but UCSF efficiency makes that cost lower, achieving the same overhead ratio but either a lower grant cost or more researcher stipend?
Plenty of actual research costs count as overhead to avoid the need to hire an army of accountants to allocate every single bit of spend.
For example, the electricity costs of the lab in which the research is run would typically be paid for by the university and would be considered overhead. It's not "administrative bloat". Most of the particularly gross administrative bloat is on the undergraduate side of things where higher tuition costs have paid for more "activities".
Note that the institution I used as an example doesn't even have undergrads. It is not using NIH grants to cross-subsidize a college. Medical research is the only thing they do. And they are the #2 recipient of grants, after Johns Hopkins.
F&A rates (facilities and administration, “indirects”) are subject to negotiation every 4 (IIRC) years, where those costs are accounted for (perhaps not well enough, but that is a separate point). The administrative component of F&A been capped at 26% for years and R1 universities are maxed out, so the negotiations are over the facilities component.
You can know what the research organization costs as a whole; and you can know what's "worth" charging to individual projects. The rest is indirect costs, which you can measure and use this data when negotiating indirect cost reimbursement with NSF or NIH.
>Could it be that these charges against grants are used to fund students in other subject areas where grants are not available?
No, PhD students from areas where funding is not available are required to teach. The university pays them for the teaching. Considerably less than what they would have had to pay someone who teaches for a living.
It’s worth clarifying that the 59% overhead rate doesnt mean 59% of the funds go to overhead. If you have a $1m grant, you add on $590k for overhead. Then the total grant is $1.59m, so actually 37% of the total funds are for overhead.
Take the army for example, it's estimated 30~40% of the workforce is dedicated to logistics. This is equivalent to 42~66% overhead (in the same sense that overhead is discussed in the context of academia, as +% cost) if you were to count only combat personnel as the direct costs.
This is was universities do, they only count research expenses as the direct costs. Yet it's quite obvious a university can't run just on scientists.
This 59% overhead is equivalent to 37% of the expenses. So, unlike you, I'm positively surprised that 63% of expenses go directly to the core mission with only 37% "waste" (which is necessary to ensure the scientists can actually work, and work efficiently).
I’m not accusing any particular organization of fraud. I am rejecting the notion that just because one institution historically receives funds that those funds were put to good use.
I understand how grants and overhead rates work. It’s an embarrassment.
"you're not literally saying fraud, but you're also not NOT saying it's fraud"
Serious accusations need serious evidence. I'm not a fan of this sowing of doubt without a solid basis to back it up. That's very much the DOGE modus operandi, and it's a lazy and dangerous form of argumentation. I'll call it out wherever I can.
You're just saying "thing bad" and expecting agreement without putting any legwork in. The onus is on the accuser, not the accused.
> The Salk Institute's overhead rate, IIRC, is 90%. Yet, they keep getting funds, so they're doing something right.
I do not agree with this statement. Take from that what you will.
> The onus is on the accuser, not the accused.
In criminal law I agree. When it comes to budgeting I do not. The onus is on every program to prove every year that they’re worth funding. I don’t accept the notion that just because something was funded in the past that it was wise then and that it’s wise to continue to fund.
So when someone says “this org has a 90% indirect cost rate and keeps getting funded” I do not think “they must be doing something right”. I instead think “wow they better have a frickin spectacular argument as to how that is possibly justifiable, and I’d bet $3.50 they don’t”.
They could have more grad students if they reliably graduated them with a PhD in four years. I was once a lab tech for two grad students that had been there 11 and 13 years respectively.
Thats insane. In experimental science there is actually an incentive for the PI to keep the grad student around (assuming they're productive) because their training is a sunk cost but its very hard to justify more than 7-8 years.
1990s was the time that the AIDS epidemic took hold. Lots of needle-sharing addicts died, reducing the number of crimes needed to support their habits. AIDS also affected the prison population, further reducing the number of active criminals.
The department store was a place where women could escape their houses or apartments and spend time in surroundings That were much more expensive and attractive. Marble floors, decorated walls, mirrors and glass, finely finished display cases, and quality goods all contributed to a pleasant environment. Seasonal changes in decoration and products kept the experience fresh. The escapist experience was paid for by pricing the goods to cover the overheads.
When the department store migrated to being the anchor store at the suburban mall, its role as an escapist haven was diluted by the mall's amenities and by the proliferation of boutique shops lining the corridors.
On-line shopping now offers a wider selection of goods at lower prices. And social media offers an alternative mechanism to escape their present reality.
> The department store was a place where women could escape their houses or apartments and spend time in surroundings That were much more expensive and attractive. Marble floors, decorated walls, mirrors and glass, finely finished display cases, and quality goods all contributed to a pleasant environment. Seasonal changes in decoration and products kept the experience fresh.
I guess it was that way at some time in upscale neighborhoods, but what I remember as a kid is linoleum and cheap worn out commercial carpet floors, the stink of cleaning chemicals, low ceilings and dim light, and unceasing elevator music. It nonetheless kept women zoned out ambling through the halls for hours. John Romero got it right when he portrayed it in his zombie movies.
The low energy music, the spread out merchandise, the allure of "sales" that could be down the next aisle, no visible windows to the outside... these stores were engineered to get people zoned out and lose track of time. Like casinos with less bells and flashing lights.
The benefits of a research project as described to the funding authority and the benefits of a research project after it is done are often not the same.
See page 10, Figure 5: Price of paper/CPI in England, 1356-1869, and the Netherlands, 1450-1800.
This shows that the price of paper declined exponentially from 1350 to 1650. The decrease in the cost of paper made printing economical. Printing on vellum was not. They point out that each Gutenberg Bible printed on vellum required the skins of 300 sheep.
>They point out that each Gutenberg Bible printed on vellum required the skins of 300 sheep.
Even that doesn't say much about its value. I have no idea the scale of the sheep industry in these times. Presumably people were growing up sheep to make use of the entire sheep, not merely to create bibles. It might be better to consider what that sheep leather might have been used for otherwise if it didn't have a buyer in the form of a printing press.
The causality can easily run in the other direction, with the desire to print things pushing people into using cheap, shoddy paper instead of quality vellum.
The cost of the hides and the labor of turning them into vellum dominated the labor of inscribing a page. Printing would not decrease the cost of finished pages, especially for small numbers of copies.
The decrease in the cost of paper resulted from the increased availability of linen rags, use of wind power in production of rag pulp, and the larger volumes of rag paper used for commercial records and correspondence before the invention of printing.
Your logic is that the drive to produce more books wouldn't lead to using cheaper paper... because paper was, at the beginning of the process, extremely expensive?
It's not clear whether it is desirable to build a new Göttingen. Communications and travel may have largely obviated the advantages of assembling brilliant minds in one place, other than temporarily for conferences and visits.
"Monumental Proof Settles Geometric Langlands Conjecture"
https://www.quantamagazine.org/monumental-proof-settles-geom...
is an interesting article that describes the multi-year effort to obtain this important result. The final proof has nine authors affiliated with seven institutions: Denis Gaitsgory, Max Planck Institute; Sam Raskin and Joakim Færgeman, Yale University; Dima Arinkin, University of Wisconsin; Nick Rozenblyum, University of Toronto; Dario Beraldo, University College London; Lin Chen, Tsinghua University; Justin Campbell and Kevin Lin, University of Chicago.
If I remember right Deep Work by Cal Newport talks about Bell Labs and how that came to be such an epicentre of invention back in the day. The author reckons that the architecture of the building had a lot to do with it. It allowed the employees to do uninterrupted, concentrated work when needed but then when exiting their offices the design of the building facilitated lots of chance encounters with other people in other departments which meant a lot of exchanging of ideas and knowledge. You don't get this kind of "hive mind" effect with remote work.
> Communications and travel may have largely obviated the advantages of assembling brilliant minds in one place, other than temporarily for conferences and visits.
Lmao. Only someone who has never been a part of a high performance team with excellent mentors available in person throughout the day, can possibly believe this.
Alternatively, try this sentence: "Communications and travel may have largely obviated the advantages of a married couple living together in one place to raise a family."
I have to agree, high performance teams need to be together in proximity to do world changing stuff (or maybe in AR worlds that are as natural as real worlds in the future). Remote work is for big standard work that doesn’t require too much innovation or if 1 person is capable of carrying the team. Yet to see counter examples to this
The analogy is a bit hyperbolic but not off base at all.
In my field, there is a lot of discussion over whether conferences should go virtual to ensure "equity", "inclusion", and all that; and to save pollution and CO2 from plane travel. According to the defenders of that take, current technology makes meeting in person totally unnecessary and online conferences can replace physical ones just fine. But curiously enough, everyone who I've seen defend that position are at elite universities in places like California where they have a high concentration of top figures in the field within a short drive. Almost no one from more remote areas (like myself) defends that... largely because many of us have stories where those polluting plane tickets helped us connect with relevant people in the field, learn and boost our career.
But back to the marriage analogy... when those people from top universities make that comment, I usually tell them that if they are so interested in equity, diversity, not polluting and all that jazz, and since according to them online interaction is enough and getting together physically is just a luxury, they should then take remote PhD students (which would even let them select from a larger pool!). It's an obvious conclusion of their position, right?
There's truth in both. For countries with little research funding (Eastern Europe in my case), travelling to top conferences is often prohibitively expensive. Top AI conferences are regularly in places like Hawaii. This excludes many regions and those researchers must submit to lower tier conferences, or - as already commonly happens - they have to move to a richer country and do research there.
Remote conferences and lower publishing fees definitely help people at these underfunded places. But it is true that being there in person is still much more valuable. Informal face to face interaction at posters, joint dinners between different research groups, mingling during coffee breaks etc. are not replaceable by a Zoom Q&A.
