[Goose]

I would imagine that those putting in the hours don't count as overhead, they would be direct costs. It's the buildings (well depreciation), the management, etc that count as overhead. Shouldn't hurt those who are doing the research at all. I could be wrong, but that's how it's accounted for in the manufacturing world, and according to Google AI, in research companies, too.

I can assure you that is NOT how it is done when dealing with the government. I have lots of direct experience with DoD and some spousal experience with NIH (and CIRM). Direct costs, overhead costs, and G&A costs are separate categories that have different accounting rules for government contracts. Your negotiated rates make a very big difference in what compensation you get. It is a game-theory operation. As the business mix varies,  companies move things from direct to indirect and visa versa. Being a "low overhead" operation is stupid if you can negotiate a high one.

At Stanford I am pretty sure lots of research functions  ("core services") are overhead, e.g. the technicians that run lots of the lab equipment do not charge direct, the glassware washers do not charge direct, etc. Simpler for everybody that way. The timecard accounting for making these services account for every 15 minutes of their time by exactly which (of the multitude) of NIH grants they are supporting at that moment would be a nightmare.

FWIW, CIRM has an overhead cap of 20%. They also allow a "facilities fee", which is classified as a direct cost(!). The government (even State government) makes its own accounting rules. This cost is calculated as a percentage of the direct grant value. Sure looks like overhead to me :-). Google "CIRM grant facilities fee". Google AI does a great job in this case.  Note that the profit vs. non-profit status makes a very large difference in how the facilities fee is calculated. After that is done, I am guessing Stanford comes close to its NIH overhead :-).

[Goose]

I would imagine that those putting in the hours don't count as overhead, they would be direct costs. It's the buildings (well depreciation), the management, etc that count as overhead. Shouldn't hurt those who are doing the research at all. I could be wrong, but that's how it's accounted for in the manufacturing world, and according to Google AI, in research companies, too.

I can assure you that is NOT how it is done when dealing with the government. I have lots of direct experience with DoD and some spousal experience with NIH (and CIRM). Direct costs, overhead costs, and G&A costs are separate categories that have different accounting rules for government contracts. Your negotiated rates make a very big difference in what compensation you get. It is a game-theory operation. As the business mix varies,  companies move things from direct to indirect and visa versa. Being a "low overhead" operation is stupid if you can negotiate a high one.

At Stanford I am pretty sure lots of research functions  ("core services") are overhead, e.g. the technicians that run lots of the lab equipment do not charge direct, the glassware washers do not charge direct, etc. Simpler for everybody that way. The timecard accounting for making these services account for every 15 minutes of their time by exactly which (of the multitude) of NIH grants they are supporting at that moment would be a nightmare.

FWIW, CIRM has an overhead cap of 20%. They also allow a "facilities fee", which is classified as a direct cost(!). The government (even State government) makes its own accounting rules. This cost is calculated as a percentage of the direct grant value. Sure looks like overhead to me :-). Google "CIRM grant facilities fee". Google AI does a great job in this case.  Note that the profit vs. non-profit status makes a very large difference in how the facilities fee is calculated. After that is done, I am guessing Stanford comes close to its NIH overhead :-).