The latest proposals for assessing research should be an improvement on the existing system, although they are hardly as radical as Sir Gareth Roberts intended when he set out on his quest for reform. Much time and expense will be saved by making alternative arrangements for institutions and departments with little hope of success, and a requirement to enter at least 80 per cent of researchers for assessment should produce a more consistent picture of standards. But the problems associated with the last research assessment exercise were more a matter of false expectations than of faulty methodology. Millions of pounds had been invested in (largely successful) attempts to improve research that, with hindsight, was never going to reach the standards applied after the event.
There is an argument that the RAE has had its day - essential at first to raise standards and concentrate excellence, justifiable as assessment was refined to ensure that true quality was identified, but redundant now that a pattern has been established. At least some of the improvements registered in 2001 can be attributed to grade inflation or universities learning to play the game. With so much money attached to the outcome, the system is bound to distort behaviour to such an extent that there have to be good reasons to allow it to continue.
Perhaps this time there are. Although even by 2007 the separation of teaching and research will be well advanced, the next RAE will offer one last chance for many departments to show that their future should not be in teaching alone. After that, it will be difficult for those who fail to break into the "world-class" categories to compete for traditional sources of research funding. Before then, however, the funding councils and the government must settle an increasingly undignified row over 6* ratings.
Their imposition has made a mockery of the independence of the Higher Education Funding Council for England and has produced a cobbled-together system that bestows riches partly on the basis of work submitted almost a decade ago. Whether there can be a fair method of delivering ministers' desire for yet more concentration at this stage in the cycle must be doubtful, but the current one certainly is not it.