Funding HQPeeps: "Fellowships" vs. The Grant

Some wacko/Nobelist (Roald Hoffman) is suggesting that American labs should be banned from funding graduate students through the PI's operating grant. Apparently, this is how the vast majority of students are funded in the States and I'm pretty sure the same is true in Canada (ie their salaries are paid through the lab's grants as opposed to an independent fellowship awarded directly to the student.) This idea has some appeal:

"The real power of an individual fellowship is that it empowers a young scientist to act in a more independent manner, on something creative and for which they have a passion," says Thomas Cech... "And that's what science is really about." Under the current system, he says, "a graduate student is told, ‘Do experiment 2a because it's in our grant.’ That turns the student into a pair of hands. So I think a shift to fellowships would be an excellent idea."

I like what Cech is saying here. The same thought occurred to me, except I was looking at POST-DOC, rather than graduate student funding. At least here in Canada, even post-docs, who are supposed to be even further along in developing their independence, are unable to get truly independent funding at all, even through external fellowships. So even if all HQPs, grad students and post-docs alike were funded through external fellowships, Cech's vision still wouldn't be assured.

The problem is that although the would-be post-doc can apply for "independent" funding directly to agencies he/she cannot do so without a PI co-applicant. So before applying for funding, would-be post-doc must meet with PI and discolses his/her interests and ideas, gain a conditional position in PI's lab, identify a putative project that fits into PI's interests and grant funding, write up a proposed project in collaboration with PI and likely write a grant proposal or two for PI pertaining to said project. So much for independence.

On the other hand, true independent funding would consist of the following: Would-be post-doc or graduate student concieves of an interesting problem and possible approaches. Writes up proposal and submits along with CV etc to funding agency. Funding agency awards fellowships to promising candidates. Awardees contacts existant labs..."I have funding to carry out the following research project. Are you interesting in hosting?"

This scheme is ultimately the most liberating for the trainee. The only real potential flaw I can see is that in theory, trainees could be granted salaries to conduct a proposed project for which the proper infrastructure does not already exist. However this could be easily overcome by properly evaluating the feasabilities of proposals.

Anyhow there's some interesting argument presented in the Science article on which is better, grant-based or independent fellowship-based funding of students. I'm not sure if either is better, actually both setups can probably work. However it's clear that the current grant-based trainee-funding regime empowers PIs at the expense of grad students and post-docs. It gives them access to a much larger putative work force, allows them to hire whoever they want and as many students/post-docs as their grant funding allows (as opposed to who and how many they can properly train). At the same time it generally tends to screw over the individual trainee by dimishing their intellectual autonomy, decreasing the availability of scientific mentorship, creating a crowded and more cut-throat and overall shittier research environment, and flooding the job market with an excess of jobless PhDs. Conversely, independent funding schemes would tend take power away from PIs and put it back in the hands of trainees to pursue their own ideas in an environment that favors quality of work over quantity. Anyone see this happening anytime soon?

Of course there are exceptions. For example, grad students who get funded through PI grants can possibly focus more directly on their research because they don't have to waste as much time filing fellowship applications or engage in meaningless CV-building arts and crafts. So long s long as they're happy with what PI decides they're working on and what training PI decides to offer, it's a sweet deal.

Overall my opinion would probably be that we don't necessarily need more independent funding for nascent grad students...what the hell do they know anyways? But truly independent fellowships for post-docs would be a great option.

Any other perspectives?