![[Home]](http://www.bayleshanks.com/cartoonbayle.png)
In science, currently, professors have to continually apply for grants in order to continue working. In some disciplines, the grant review committees assess only the quality of the proposal, rather than the interest of the topic, but in many disciplines, they do both. I am arguing against both, but particularly the latter.
Assume all professors are equally smart and well-informed. Let's say that there are three potential lines of inquiry available. Say that 1% of professors think a certain line of inquiry is a good idea, and 9% think a second line is a good idea, and 90% think a third line is a good idea. Assuming that review committee composition mirrors these proportions, almost 100% of funding will go to the third line if inquiry, because 90% of the committee will vote for (I use "vote" in a general way, to mean "express a preference which is taken account of via some procedure") projects along line 3 (assuming that the committees do not use a proportional representation-ish system of budgeting, but currently they usually don't).
Now, I think that's suboptimal. All of the professors are equally smart, hence before you knew which line of research they chose, you would have said that they are all equally likely to be right, so if you had to bet, you probably wouldn't assign a near 100% probability that line 3 is best; you'd probably assign about a 90% chance. In other words, the probability that a given line of research is best is the same as the proportion of professors who prefer it.
Now, should we allocate money between lines of inquiry proportionately to their chance of success, or should we give all the money to the most likely line, or something else? This is a difficult question, because it depends on things like, are there diminishing returns by spending more and more money on one line of inquiry? How quickly and how reliably do you learn when an initially promising line of inquiry is wrong? But I think few people would argue that 100% of the money should go to the most likely candidate line, which is probably a fair approximation of what happens today. Instead, I'll assume that the correct thing to do is at the other extreme: allocate money proportionately to the liklihood of success.
So, what we need is a funding system that assigns money to lines of inquiry in proportion to the number of professors who like it.
One such system would be to introduce proportional budgeting processes into the review committees.
But an even simpler system is to simply give each scientist the same amount of money. I call this the "uniform funding" system.
Two solutions. First, scientists can band together into groups and pool funding.
Second, the uniform funding system could be used to distribute most of the money, but not all of it. The rest of the money could be distributed via a grant system similar to today's.
If you think this is important, than instead of reviewing projects, committees could review professors, and some of the funding would be earmarked to give only to the highly-ranked professors.
Instead of or in addition to using committees, one could use a reputation economy procedure such as a "reputation budget" currency combined with a dominant eigenvalue distribution method ... but describing that is beyond the scope of this page.
See the answer to the previous question.
Also, note that additional institutional mechanisms could be set up to catch really flagrant abusers (i.e. people who basically just don't work at all and try to collect the uniform distribution money anyway).
I don't think this would be a very big problem because everyone I know in academia is so interested in their work that they wouldn't just go and play golf rather than work even if they could. Many professors work after retirement, and I've known at least one professor who worked and taught classes at his home institution right through his sabbatical year (he said he took the sabbatical just so that he wouldn't have to serve on committees that year!) So I doubt that removing financial incentive would cause the work to slow down immensely. But it may have some effect. However, is that effect worth the cost? See below.
I think it's a good idea to have some diversity in the funding model. If I were dictator, I would distribute 70% of the money according to the uniform system, 15% going via a system that gives more money to high performers (half of the money distributed by committee, and half by reputation budget w/ dominant eigenvalue, except that the lowest half of the contestents would not get any of the reputation budget-derived money), and 15% via a traditional project-based grant system for large projects (note that that 15% to high-performers skews income more than you might think, since its distribution is concentrated in a small fraction of the population; for example, even if you just distribute that 15% evenly among the top half of researchers, then better-than-average researchers get over 35% more money than worse-than-average ones ((1 + .15) / (1 - .15)); if 20% were reserved for high-performers, then a high performer would get 50% more than a low performer) (interesting side-note: in terms of incentives provided to individual researchers, this is somewhat analogous to a corporation which spends 70% of its payroll on salary and benefits, 15% on bonuses, and 15% on hiring consultants (where the consultants are employees who are so good that they have spare time to consult) (if you redefine "payroll" to include consultants) -- the large proportion of salary can be explained by a recognition that in this case, the objective metrics that are used to allocate the bonuses are very unreliable).
However, I argue against devoting more than half of the money to these more complex variations. Why? These variations require grant-writing and review, and it is extremely desirable for most professors to be able to spend very little time on either of these activities. If only a small fraction of the money is distributed via review committees, then most people can go most of their careers without paying much attention to grants and reviews.
But, you say, I expect that review committees can direct the money more efficiently than uniform distribution, and I expect that by providing more review, professors will be incentivized to work harder.
My retort is this. If grants are important, everyone will feel that they have to be involved in the grant system to stay competitive, and you get the current system, in which a significant fraction of the total pool of time of professional researchers is spent on grant writing and on reviewing grants. I don't know what the fraction of time is, but I'd venture that it's over 10%, maybe as high as 20%. In addition, there are more subtle costs in terms of people's uncertainty about future funding causing them to spend their time in less productive ways than they could if they had steady, reliable funding. In any case, I'd venture to say that about 20% of productivity is lost due to the grant system.
Above, I argued that a grant/review system leads to inefficiency by underallocating money to unpopular approaches. But, even if you think a grant/review committee system leads to more efficiency (both by allocating money better, and by incentivizing professors), you have to weigh that against the 20% overhead cost. Do you really think the grant/review committee system boosts efficiency more than 20%?
