In the good old days, this was a typical scenario for NSF grants, indicated by time (t) in months:
t = 0: proposal submitted on or before the proposal deadline. (I was just trying to remember if I have ever submitted a proposal before a deadline, and I don't think I have.)
t = 3.5-4: panel met, discussed mail reviews, ranked proposals
t = 4-4.5: program directors made decisions. Those PIs who were definitely funded and those who were definitely not funded got the news right away; those who might be funded waited a bit longer.
t = 5-6: proposals recommended for funding worked their way up through the system, but it didn't take long to get the final award letter and for the funds to be transferred to the university and for the university to assign the grant an ID number, making the grant active.
Back in those glorious days of yore, when filling in the cover page, I would indicate a preferred start date 6 months after the proposal due date, and this was quite reasonable. A new grant was typically good to go after t = 6 from time of proposal submission.
This was an excellent system because you knew that if your grant was funded, you could start paying a graduate student RA in the next academic term. This was very useful for making a quick start with the research, for trying to optimize correspondence of graduate students and grants, and for research planning in general (e.g., if you get/don't get a grant, this affects your plans for the next proposal deadline).
Now the system seems to be more like this:
t = 0: proposal submitted on or before the proposal deadline.
t = 3.5-4: panel meets, discusses mail reviews, ranks proposals
t = 4-8+: program director makes decisions, contacts PIs
t = 6-10+: proposals recommended for funding work their way up through the system, and eventually result in the formal award letter, which will be followed at some point by transfer of funds to the university.
And then.. there is another wait to get the university to assign a magic number to the grant so that it can be used. This delay is beyond the control of NSF, but it is one more delay on top of the other delays, making for a considerable gap in time from proposal submission to effective grant start date. I specify effective grant start date because now the start date indicated by NSF may be a couple of months before the university completes its paper work and recognizes the grant in its grants management system. The grant may technically exist for a while before the PI can use it.
It is sometimes possible to start using a grant that has been promised but that hasn't quite worked its way through the system to the final stage yet, but to do this you have to have a slush fund (not another NSF grant) as back-up.
Why has this all become more complicated and slow? Is it because budgets have been slashed and there are fewer staff handling more responsibilities? I have a hypothesis, but it actually involves the opposite of this explanation.
I know that there are all sorts of considerations involving oversight of every step of the system and oversight of the overseers and so on and this all takes time to make sure no one is doing anything unethical or illegal of unwise, but I wish we could go back to the 6 month proposal-to-grant time gap. Gaps that are considerably longer than that seriously interfere with a PI's ability to assemble an excellent research team and do exciting research in a timely way.
I am of course happy to get grants at all, so it might seem ungrateful to complain about a delay of a few months, but I have found that the delay has rather significantly affected my research program.
2 years ago