May I propose a new parlour game for Christmas - spot the unlucky research-assessment-exercise gambler ("RAE league table of excellence", THES , December 14)? In economics, I'd guess Oxford and Cambridge could have submitted everyone and got a 5. Instead, each unit left out a tenth of staff, gambling that this would bring a 5* and 10 per cent more money. They were wrong, stayed at 5 and got 10 per cent less than a full submission would have given. But the elite has it relatively easy. Submitting one researcher too many could cost a third of your funding when you are gambling lower down the scale.
Gambling for public money is absurd. Instead, the RAE could count the number of "national" and "international" researchers in each unit. Allocate, say, £10,000 per "national" researcher, and 2.5 times as much per "international", varying the first figure by subject, and you would get something close to the current funding scale but with no cliff edges.
Moreover, universities could not lose money by submitting extra researchers, so would not have to waste hours agonising over whom to submit.
Department of economics
University of Hull