An interesting question...

Around a month ago, Drew Gilpin Faust of Harvard was quoted by MSNBC as saying: (via CosmicVariance )

One thing we all must worry about — I certainly do — is the federal support for scientific research. And are we all going to be chasing ncreasingly scarce dollars?” says Drew Gilpin Faust, Harvard’s new president. Not that Faust seems worried about Harvard or other top-tier esearch schools.

“They’re going to be—we hope, we trust, we assume—the survivors in this race,” she says. As for the many lesser universities likely to lose market share, she adds, they would be wise to really emphasize social science or humanities and have science endeavors that are not as ambitious” as those of Harvard and its peers.

Five years ago I would have found such a statement obnoxious, to say the least.  Today my opinion is that she has a point.  I feel that government funding should go only to the top 15 universities in a particular field.  Here's one reason for my change of mind. 

I have worked on a multi million dollar DOE project for several years at  a not-quite-top university. I think it was mostly a colossal waste of taxpayer oney (even though it fed and clothed me). Most of the money would probably have been better spent at a university that could attract more roductive researchers, i.e., a top 15 university.

On the other hand, it may be that the number of truly productive researchers is severely limited. In that case if more money goes to the top universities all e will have is more mediocre researchers at those places. But even mediocre researchers can reach greater heights if they are exposed to great eople (who presumably inhabit the top universities).

I think scientific activity will be more productive if it is concentrated at a few places rather than dispersed. We might be late to rrive at a few unusual ideas but in the long run we will arrive at those ideas. And, we will have fewer unemployed/unemployable PhDs :)

 About 5-6 years ago I'd have found Drew Gilpin Faust's statement entirely correct. Today, I don't think this is so. For a variety of reasons.

Assume for example that Researcher A is a great, brilliant. To develop his/her ideas (s)he needs money for himself and for his/her graduate student, and for research apparatus. Now, after a certain point, lets say 200K (salary) a year, there is no difference in the research productivity of the scholar. Let's also say that after 10 (graduate students + postdocs) there is no increase in research productivity. Let's denote this level of funding by L1.

 That doesn't mean that the funding agency should fund only the top 15 at level L1. The funding agency should do this and more. The reason being that it is impossible to predict where the next big idea, the next big theorem could come from. I'm sure you know of Ramanujan, Gigori Perelman. In my opinion, similar things happen, more commonly on a smaller scale. Some people work better in the environment of a big university, some people don't.  e.g. Mario Capechi (sp?), this years Nobel prize winner, was probably from your university. 

Finally we should note that researchers aren't machines whose output increases linearly or superlinearly with resources. There is a point of diminishing returns. 

Finally, on a personal note: don't be so hard on yourself. 

 -Amit