Hacker News Reader

Comment by avsteele | Hacker News Reader

This a decent write up. The SBIR budgets wind up looking similar.

I can see how the indirect costs ("overhead") seem like ripoff to a professor, but even if you are running a private business having an indirect cost rate of 50% is considered low-to-normal. At least where I am: rent, utilities, cleaning, bathrooms, etc... are just expensive.

View on HN

The problem with indirect costs is that it's well documented that the actual costs to the university are less than the amount received from grants. So the university is profiting in a very real sense. This has been written about by former federal agency heads and so forth and so on. Comparing costs to private businesses isn't quite right because some of those costs aren't actually borne by the university, they're done in house, they otherwise get discounts, or they're already covered by other income streams (to a certain extent you get into double charging for the same thing really fast).

You can intuit it just by how it's distributed, as a percent of grant funding. This really doesn't make sense, it's not tied to a line item budget or anything.

The problem with this is that it creates a perverse incentive for faculty to bring in large grants monetarily speaking, rather than to do good science per se. It's one of the reasons for the pyramid scheme we've seen over the last couple of decades at least, and part of the problem with how academics is incentivized.

Indirect costs should be cut completely IMHO (speaking as someone in academics, who has had tenure, and been involved in grants). Yes it will be painful but maybe then there can be a real conversation about what's going on in academics and how people are rewarded. Maybe things should be line item budgeted, maybe universities should just receive a lump sum in exchange for certain practices, maybe states should actually fund public universities in line with what they want them to do, I have no idea but right now the system is broken.

I'd be interested to see the data for unis showing their costs are lower. I doubt it.

For private companies, and an SBIR ,you have to establish an indirect rate based on your actual costs. The rate is the ratio of all indirect costs for the company vs the direct costs for the project (salary times the % of the person's time). It's a bit of work the first time you have to do it.

I'd be surprised if the NSF lets the uni just put whatever they want down and it isn't based on submitted budget data.

At the US university where I worked the indirect rate for federal grants was the same across the entire university, and I think it was slightly more than the 53% OP pays. I don't know where the number came from. Most of our industrial grants had language in the contract that said they would pay whatever rate the government was paying.

No data for it, but several times we've told companies if they think it can be done cheaper elsewhere for the same quality that they should do that.

We've had no takers.

Heck, one of my major sources of consulting income is "Checking the math a private industry research group did."

At my university when you look at it as a whole, we barely break even on indirects.

While there is a lot of need to have a conversation about how the system currently functions, I'd suggest cutting indirects completely as step one is akin to suggesting flooding the room with carbon monoxide until we can sort out that nasty cough you have.

Why doesn’t the university apply for indirects via a separate grant?

It seems weird to tuck it into the research grants. More professors with a research grant do not automatically translate to more indirect costs (e.g. the building has to be cleaned, whether a room is occupied or not)

It is at least intended to reflect the cost of the grant. The building might need to be cleaned, but one could expect things like the library, a laboratory, support folks in IT and grant administration, the IRB, etc. to go up.

In addition there is a world wide trend that the ratio of administrative vs academic staff keeps on increasing, in other words less and less people bring in the money (via teaching/research) that sustains the university. Hence the overheads go up, because those people need to be paid.

Even more infuriating is that when processes get reformed/changed it almost always ends up such that the work of admin gets easier at the cost of academic workload (because academics typically don't have set hours and just do the work until its done). To give 2 examples, at a previous uni there was always a lot of buerocracy about putting marks into the system (lots of paperwork, especially if there was some accidental mistake so a mark had to get changed). So during the end of the semester the uni forbid admin staff to do overtime for cost reasons, however the work still had to get done, so academics were doing the admin work as they were not on a fixed working time.

Second, at my current uni, when getting travel reimbursements, we are required to fill out an online form, scan all receipts and print out a generated form and send the form and original receipts to finance. So academics end up spending time scanning and printing documents that could just as well be done by a person at the finance.

> The problem with this is that it creates a perverse incentive for faculty to bring in large grants monetarily speaking, rather than to do good science per se.

And it creates a straight-forward incentive to do good science to so grantors give you big grants.

I think one reason indirect costs get a bad rap is there are supplies that can only be paid for through overhead for example ... pens. Depending on the institute, getting anything back in terms of support can be tricky or down right impossible. It was so bad in one place, that we would all go out of pocket for notebooks and pens because the indirect portion was a black hole of nonsupport.

I think this is the biggest thing. Universities have to report all of these things separately - private industry gets to tuck it into things. Having seen the comparative rates for private research companies...they're not cheaper than universities. Not by a long shot.

IMHO it's outrage-bait from an AI researcher who is likely oblivious to the costs of the computing infrastructure they use and thinks the overhead being charged on top of the grant award is outrageous because they don't have lab space and such.

Sadly, most researchers have zero appreciation for how much computing infrastructure costs, and how they think it's perfectly fine to store data for a several million dollar research grant, that involves multiple people's academic careers, on a USB hard drive they got at Best Buy because buying a proper server etc would cost too much.

You can see their attitude in how outraged they seem to be about having to pay a grad student $46k a year...

I only paid my students $26,000 a year (listed as $2200 a month in the post). I wish I could have paid them many more times that like they deserve!

If you are a CS researcher, the chances are the computing infrastructure you need is very cheap. If you only develop algorithms and software, your computing requirements are probably an order of magnitude smaller than for people who use the stuff you developed in their research. And even your data is often just something you downloaded from a public source, because you needed something with the right combinatorial and statistical properties.

they don't pay the grad student $46k a year. they pay them $24k a year (or more likely $20k for 9 months), and then the university makes up a number for "tuition" and takes an extra $15k arbitrarily.

Reply on news.ycombinator.com