Idea: A better way to allocate grants for research
by Alan Cohen
There is a paradox that most granting agencies face when they fund research: Researchers have a strong incentive to ask for as much money as possible even if they don’t really need it. One reason is that researchers are evaluated by both their institutions and their peers based on the number of grant dollars obtained (rather than on the quality of their research or their productivity). Another reason is that it’s always better to have a little too much money in the budget than not enough – we might as well estimate high.
These incentives are at work all the time in subtle ways. They mean that researchers ask for more money than they need. There is thus a slow inflation in the accepted “cost of doing business” and in the amounts viewed as normal. There is no reason at all for researchers to try to get by on a shoestring. Perhaps most insidiously, there is an incentive for young researchers to choose research topics that require more money for the same amount of productivity. Several years ago at the University of Michigan, I heard that one of the deans wanted to hire fewer faculty in ecology and evolution and more in lab-based biology because the latter brought in larger grants.* I often see colleagues in developing countries finding cheap and inventive ways to do the same things we do here, just on a fraction of our budgets.
In addition, most granting agencies demand detailed budgets for submitted projects. These budgets require lots of time to prepare, lots of time for peers to review, and are mostly made up anyway – people say in the budget what they need to say to get funded, knowing they can use the money as they wish later. I have even been told by CIHR representatives that I should fudge my budget because it is the best way to get funded!
One reason these problems persist is that in most cases successful grants are chosen with regard to quality but without regard to value. In other words, if there is $1 million left in the budget and a $1 million grant is ranked just above 10 grants that are $100,000, the $1 million grant will be funded and none of the next 10 will be, even if the total value of the next ten projects is 9.5 times that of the $1 million project.
I’ve just devised a system that I think elegantly solves this problem and substantially reduces paperwork. Here is a summary:
1) For competitions, the funding agency decides a priori on a number of funding tiers (say 7, ranging from $80,000/year to $1 million per year). Each tier has a number of slots available based on the total budget, with many slots available at the lowest tiers and very few at the highest.
2) Each researcher does not fill out a budget, but simply selects a tier he or she wishes to be considered for. If the grant is awarded, this is the amount that will be received – no more, no less.
3) All applications are ranked based on quality, without regard to the budget tier chosen.
4) Funded applications are chosen starting at the lowest tier and choosing the best candidates who applied for that tier, until the number of slots allotted for that tier are filled.
5) At each higher tier, any applications not chosen for the lower tiers are retained in the competition, and will still be funded at their requested (lower) tier if they rank sufficiently high among the higher-tier applications. In other words, at each tier the highest-ranked applications are funded, including those at that tier and those not selected at lower tiers.
6) Any remaining funds (due, for example, to lower-tier candidates winning higher-tier slots) can be given to the remaining candidates in order of their rank, irrespective of tier.
This system has a number of clear advantages:
1) It removes a huge amount of work from both writing and reviewing grants – everything related to budgets, which are usually fudged anyway.
2) It gives a strong incentive to researchers to conduct their research on the smallest budget possible, since the chances of obtaining a grant are much larger for those with smaller budgets.
3) It gives a way to fund larger-budget projects or programs when justified, and lets the competition incentivize the researchers to make the determination about necessity.
4) It gives an incentive to young researchers to choose relatively cost-effective fields of study, improving the cost efficiency of research for decades to come.
5) It rewards overall candidate quality and frugality simultaneously, in such a way that budget does not become a perverse incentive to do other than the best research.
What do you think? Would it work? What are the challenges?
*Caveat: I heard this story second-hand years ago, and may have misremembered something in the details. But there was unquestionably a perception that expensive research was prioritized by someone higher up.