Having been in the university science game for over 40 years, Eli, well actually, Ms. Rabett always gives a snort when told how everbunny is getting rich on grants. Katherine Hayhoe, having perhaps reached the end of her very long rope on this, put something up on Facebook, explaining why no bunny gets rich on grants.
There are a couple of things buried in here that need some further introspection. Eli, having had more than his share of luck, success and skill (Eli is currently available to help fine tune your grant for suitable carrots) in the grant game would like to add a few points. These are US centric, and pretty much on the physical sciences side of the house, but there are carry overs.the biggest grant I was ever awarded was the stunning amount of $1.1M USD. Stunning, that is, until we break it down. Here goes!
I wrote the grant with 4 co-principal investigators from different universities. So we divided up the money about equally, giving each of us around $220,000. Still a nice amount.
The grant was for 4 years, which meant I had $55,500 to spend each year. Still nice, right?
Then the university takes 1/3 of that for "facilities & administration" costs. F&A doesn't cover the luxuries -- my office has no window, I bought all my own furniture and computer -- but it does give us internet, electricity, and an infernal amount of paperwork. That leaves $37,000 for me to spend each year.
With that money, I pay a graduate student the princely salary of about $25,000; I pay the university their tuition, which is around $10,000; and that leaves $2,000 left over each year. Year one, I buy the student a computer; year two, I pay for them to attend a scientific conference; and years 3 and 4, I pay for us to publish one scientific paper because yes, those cost about $2,000 as well.
BOOM - that's how a scientist spends $1.1M! Surprised?
Also not a complaint. Eli is very grateful for the support he got, and in a sense for the support he didn't get when proposals were turned down, sometimes on interesting grounds. OK understanding. For what it's worth the only time the Rabett got mad is when the first reviews said do this, Eli did that, and the second round said why didn't you propose what you did the first time. Ah well.
First, almost all US academics outside of the health sciences, are on 9 month contracts. That means that their salary is paid 9 months of the year, and in the three, usually summer, months they can a) go lie on the beach b) teach summer school and make a pittance, c) do research elsewhere like in a National Lab or d) get support from grants. Monthly salary is limited per month to 1/9th of the academic year salary, what is called their institutional base salary or IBS. Department Chairs usually get an additional month salary for the abuse they take and Deans, Deanlets and higher have 12 month salaries, but their "off planet" time is limited.
Teaching loads vary from place to place and from department to department, but generally scale with the amount of funding going to a department or area. The number of slots scales not only with the number of students to be taught, but the external support the faculty bring in which is the driver for departments and deans demanding faculty write grants. Faculty can put support into their proposals for academic year teaching relief, but this is very hard to get and does not raise your IBS, just decreases your teaching load with the money going to the uni. In a few research centered places it gets really uncomfortable if a faculty member does not buy out some teaching.
In general, an NSF or NASA or DOD grant will not support more than one month summer salary, so the need for multiple projects, but you can't have three, because the agencies agree that you will probably then take vacation on their dime. Then in addition to the summer salary there is a 25 - 30% surcharge for fringe benefits. A large part of the FB in the summer is a fiction, because health care costs, among other things, are completely covered from the 9 month salary. In some cases there are also no additional retirement benefit payments. Salaries and fringe benefits are subject to the Facilities and Administrative charge, which is for such things as the cost of the library, cost of space, electric, water, etc. etc. While these used to be all over the place, they are now pretty much in the 40-60% range peaking in about the middle, so on summer salary, there is an additional charge and putting FB and F&A together the grant has to cover 180% of the cost of the summer salary. For assistant profs, this might be ~ $15K, for senior faculty up to $40K/mo
The F&A money gets sliced up. The fact that the granting agency has paid the university for use of its facilities and administrative costs, means that the recovered costs can be used for what the school wants. The money belongs to it. The University retains the major share, but portions usually are placed into separate accounts for the Dean, the Department and the faculty member. These funds are unrestricted and used for such things as start up funds for new faculty, funds for large proposals, usually for major equipment, or for conference travel support.
Post-doc salaries cost btw $45-50K per year depending and there are those who low ball. The NIH guidelines are not a bad estimate, but the true cost is an additional $40K for fringe benefits and F&A, and on top of this costs for travel to conferences and such need to be added, so $95K per postdoc is a good estimate.
Students, aka participant costs, do not usually carry fringe benefits beyond payroll taxes (~8%), but their health care is buried into tuition and fees. As Prof. Hayhoe writes, the cost of the stipend is ~25K$. Tuition is interesting. For state universities tuition for non-state residents can be much higher than for in state US citizens and foreign students all have to pay the out of state rate. Cost of tuition also varies with policy about tuition for summer semesters, but another $20K per year would not be unreasonable, and in some places up to $40K year, so a graduate student charge on a grant would be between $45K and $65K/yr, not including sundries (there are whole lists of strange fees to be paid). One place Eli knows about starts with $0 and then accelerates up to full tuition over a few years to encourage PIs to graduate their students.
As a consequence, there is a pressure to keep foreign students working as Teaching Assistants whose stipend AND tuition is paid by the university. This, of course, does wonders for how fast they graduate.
Today, with agencies insisting on open publication, between $2-4K per publication has to be allowed for and also travel to a conference, say another $3K.
Now, of course, something needs to be done, field trips, special equipment (over $5K carries no F&A) materials and supplies, and, of course, there is F&A to be paid on all this, which means a $1 nut, costs $1.50. . . which means if you want to buy something, it's better to buy the more expensive one that costs $5K than the one that costs $4K, and of course, if you are building something, don't budget it by nut and bolt, but as a whole. Also, as a rule you have to make a strong argument to buy a computer, unless it has special features necessary for the work. Lots of prose expended on that one.
So, now go forward young Rabetts and write your proposal.