Wednesday, January 19, 2011

he Long Tail of Research, Full Economic Costs and Government Cuts - Journalists see here (FIA?)

Research value in the long tail is an interesting idea - so industry and shorter term pressures on results do produce applicable ideas, but it seems to me that not only is it harder to attract such funding from industry for longer term (5-50 years) research, but that it is way way harder to predict (and to attribute/measure impact after the event) results, which is why government funds things.

So if we look at the distribution of funding over individuals in Universities (or labs that do long term research), it too is heavy/long tailed.

Now with the cuts in funding to many areas in the UK, the thing that goes first is the thing with least measureable outcomes. And money concentrates at the top of the curve.

A while back, UK universities agreed to seek Full Economic Costs for each and every thing they do, from all funding sources possible. What effect do you think this has had? Well, my guess is it has made it even harder for long term basic/foundational researchers to get funded, and easier to get money for short term people (like me:). Now, an interesting question is "how has long term research survived recently, then"?

Answer. FEC is being recovered at >100% for the people in the short term areas, and siphoned off for cross subsidy for long term. i.e. we're doing something he government used to do by stealth.

However, this is probably not legal - certainly if you bill more than 100% to EU, even if you deliver the work, you are working more hours than european law permits (almost certainly true for most UK universities, and other large state systems like the NHS, and education in general).

But worse. It is distorting, and creates an incentive for the government to go on under-funding long term work, because they don't see it vanish quite so fast as it is kept limping along under cover of succesful short term work.

Some journalist (THES maybe) should ask for the costings for some universities research projects and the charges made by PI name (under an FIA request) and see how often the time adds up to > 100%. Note, some of those PIs are supposed to be doing teaching as well. Some of them don't even actually cost any money at all, if their posts are endowed, or they are emeritus.
(If you think that the funding agencies would do this, think again - the EU and UK National agencies aren't going to be swapping data bases, for lots of ok reasons).

It is noteworthy that (rumour has it that) Intel has decided to shut down its "embedded" lablets (Seattle, Berkeley, Pittsburgh) after a successful decade of interesting research, but few results that benefited the corporation mission/goal/objectives, even if the ideas did see a lot of cool publications and may well have been useful elsewhere. And other such labs funded by other temporarily enlightened corporations may have benefitted Intel in turn . Despite its best ever financial quarter, it isn't that surprising that Intel finally called time on this disruptive experiment started so nicely by David Tennenhouse. It is the job of governments and state subsidy really, and why we charge profitable outfits like Intel tax. ANd why we give them tax breaks when they donate money to long term research.