> How should government fund basic vs. moonshot research? Incentives. If they think they are smart enough to allocate funds via a committee, they need to be smart enough to figure out how to structure incentives.I can't find the reference right now, but fairly recently there was a report on the declining ROI of NSF (or possibly NIH) research grants, in that the bias toward proposals that seem likely to produce results was foreclosing high risk high reward proposals, while the funded proposals don't actually fulfill the expectations of the selection committee (in part because what was really being selected for was the ability to write a convincing proposal).
Committees just aren't all that good at their job of picking research proposals (or researchers) that are likely to produce results that advance our knowledge.
Several proposed remedies were examined, but the one that seemed best able to overcome the issue was reserving a tranche of funding for proposals randomly selected from a pool (said pool only excluding the very worst proposals that don't meet basic eligibility requirements).
