That appears to be one of the suggestions for how to encourage more risky research in a recent article in Physics World (H/T Scientific Blogging).  Risky in this case is meant to be different, unique, cutting-edge, not high potential for harm or damage.  This is another effort to address a perennial research management question – how do you encourage enough of this kind of risky research in order to keep getting the kind of radical breakthroughs that push science and technology along.  Science is an inherently conservative (small-c) process, as it requires the agreement of the community to establish its information.  Those who take risks and fail lose a lot, as tenure committees, journal boards, and funding organizations don’t deal well with ‘failed’ experiments.

Some of the suggestions in the Physics World article are familiar with those following ‘blue sky’ research – programs or agencies focused on this kind of research, a percentage of research funding set-aside for such research, fellowships for out-of-the-box thinkers, tenure as a means of allowing experienced researchers the job security to take these risks, etc..  They all have their benefits and problems.

What was novel – at least for me – were two strategies taken from the investment realm.  One boils down to critics and supporters of new or radical ideas putting their money where their mouth is, so to speak.  You could create an investment or prediction market on a particular scientific proposition.  This assumes – and it may be a really big assumption – that the success of prediction markets in other areas (look at those involved in political races) will transfer to scientific research.  It’s also unclear if enough people would have enough interest in the research question, one way or the other, to risk enough money in the market to make this viable.  I get the impression the market wouldn’t have to be big enough to actually fund the research, since the money in the market (or most of it) would go to those who picked correctly.  The market would have to be big enough to convince some funder that the question is worth pursuing.

The other idea would be to somehow approach research funding as an investor might try and develop a mutual fund or retirement account.  Invest enough in conservative research – research that will have results as it is expanding or refining the knowledge in a field – to allow for investing in a few riskier ventures.  Unfortunately, this strategy is a bit more problematic because of how returns on scientific research investments are accounted.  Exactly how would you measure return on investment?  It makes sense to me to approach scientific research more explicitly as an investment, but there are many counting measures to address before making this work.