undefined | Better HN

0 pointsSJC_Hacker1y ago0 comments

Its "normal" because its generally accepted, not because it makes any sense.

I was part of a research lab on grants like that. We had close to $1m in total funding, on top of that indirect was like 50% (so $500k/year) We maybe had 4000 sq. foot of lab space in an old building that wasn't maintained well. We had one bathroom for each gender on the floor for the research arm of two whole medical departments. Two admins for the whole research department of 7-8 labs totallying maybe 60-70 staff.

I ran the numbers and the lab space would have maybe cost $100k/year tops (probably more like $80k, depending on quality) if we were rent out equivalent industrial office space. On top of that you have electrical, heating, telecom, at most $10k. Support services such as HR, cleaning, IT support (of which we didn't use a whole lot) could have been contracted out, at most around $20k. So there was about $350k which I figured was mostly just a subsidy and went to "administration". Not that I was philosophically opposed to it, except maybe the admin.

0 comments

dgacmu1y ago

You're actually still misunderstanding overhead a little.

Overhead isn't applied uniformly. For example, tuition for Ph.D. students isn't charged overhead, nor is (usually) equipment. So on $1m of funding, if you've got 4 Ph.D. students, that may be something like $200k/year of tuition that isn't subject to overhead. Add in another $100k of equipment and suddenly that 50% indirect cost rate is actually more like 35%, so you end up doing $1m of "work" on $1.35m of budget.

Departments often negotiate something called "overhead return", which is a way of returning a small amount of money to the individual departments -- some of this does things like supporting Ph.D. students if their advisor runs out of funds, or helping research faculty bridge short funding gaps. These things are reasonable and help the institution remain coherent through the uncertainty of grant-driven existence.

There's waste everywhere, but it's not quite as bad as it might seem without a deeper understanding of the university research funding model.

mlyle1y ago

It's also worth noting that this overhead percentage is misleading. A lot of other contexts would view $1M of work on $1.35M of budget as 25% overhead, not 35%.

SideQuark1y ago

The parts you list result in wasted research money. The system you complain about results in more R&D getting out into the world.

The money you complain about goes to run an org that has connections, does advertising, provides stable employment when grants fluctuate, has hiring and HR and payroll and a zillion other services, all making those doing the research more able to do research, and provides more channels to move results into production.

So it makes sense. You just haven’t thought through or had to perform all the pieces, so to you it doesn’t make sense.

monero-xmr1y ago

Ah yes, we need a fudge factor to subsidize a bunch of unquantifiable woo like “advertising” and “stable employment”.

I’m sick and tired of elites telling me basic business operations of profit and loss, value for money, quantifiable results are beyond my peasant brain to understand.

vkou1y ago

> Ah yes, we need a fudge factor to subsidize a bunch of unquantifiable woo like “advertising” and “stable employment”.

Try convincing the AC guys to work for parts cost + a skilled worker wage * number of hours worked, see how well that goes over. They'll laugh you out of the room, and you'll be left sitting on your ass without air.

The entire world charges overhead for work done. Most of it way more than 25% of the sticker price.

1 more reply

cozzyd1y ago

Typically overhead is only charged on a portion of expenses. In our case, anything over $5k or that is part of a "constructed equipment" over $5k (these two categories are the large majorities of expenses in our lab, as most things we buy are components of detectors we build) are overhead free. Supplies/laptops/travel/tools/business meals/inexpensive equipment do incur overhead, but the effective overhead rate is much less than the nominal one.

j / k navigate · click thread line to collapse