Penalised for making it

AFTER the emotional responses to the research assessment exercise results last week, attention shifts to the financial implications.

The quality research funding formula is not yet known. It is believed that 5-star departments will be treated the same as grade 5, there is uncertainty about grades 3a and 3b, and whether money will be allocated to grade 2. Nor do we know how much money will be available for each unit of assessment, so in the table below I have assumed the same distribution as before. To illustrate the strong case for a new formula, I show here that continued reliance on the old one means that many of the country's strongest research departments will be penalised.

Brian Fender tells us that our research is getting better. Average grades are higher if the old scoring system is used. But more staff have been assessed in most units and in more departments.

Any department's gains must be matched by another's loss. If this means that higher grades bring more money and declining grades less, it might be seen as equitable. But departments retaining the same grade or even higher could get less if their unit of assessment experienced either grade drift (the average grade was higher than in 1996 than 1992) or size drift (the average department was larger in 1996). Weighted grades (grades multiplied by staff numbers returned) are higher in 1996 in most of my sampled units of assessment.

The geography unit of assessment illustrates this: 50 departments with 696 staff were assessed in England in 1992; 1996 saw 58 with 904 staff. There was grade drift: the average unweighted grades were 2.86 in 1992 and 2.98 in 1996 (on the old five-point scale); the average weighted grades were 3.34 and 3.55 respectively. Bigger departments got better. Under the old formula I estimate that:

* the eight departments assessed in 1996 but not 1992, with at least grade 2, could share Pounds 300,000. Of the remainder, 18 will lose money and 21 will gain

* some of those with higher grades in 1996 will get more money, but one that went from 3 to 4 may lose some. Four departments went from grade 4 in 1992 to grade 5 in 1996, but only one will get a large reward (perhaps Pounds 125,000 - its staff increased from 21 to 31); the other four, whose staff increased by about one-quarter, may each get only Pounds 50,000 extra per annum

* of the four that moved from 5 to 5 star and got the accolade of England's best geography departments, one will gain a small amount (perhaps Pounds 50,000: it more than doubled the number of staff returned). Two will neither gain nor lose (although both increased their staff by 40 per cent). The other could lose as much as Pounds 200,000 (its staff increased by only one)!

* departments retaining the same grade will almost certainly lose money: all six that retained a grade 4 will lose up to 40 per cent of their QR income

Departments had to play an uncertain numbers game when preparing their RAE submissions. They guessed that to slip a grade point would be financially disastrous, and to stay on the same grade might at best be financially neutral; their only hope for monetary gain was a higher grade. They also knew that the more staff entered as research active, the better the financial outcome, if a grade point was not put at jeopardy by entering "relatively low achievers". Would a grade 5 with 20 staff bring more money that a grade 4 with 28, and would entering 25 reduce the likelihood of getting the 5? And was the more prestigious 5 better anyhow?

Those with a grade 5 already could aspire to grade 5 star, but believed there would be no increased reward. To earn more they must enter more - which might put the higher grade at risk.

Managers at Oxford and Cambridge apparently adopted different strategies. Oxford went for the grades, by excluding nearly 10 per cent of all staff; Cambridge excluded only 2 per cent, and presumably went for volume. So Oxford has the glory, but Cambridge may get the money!

The Higher Education Funding Council for England must introduce a formula that does not penalise success, otherwise it puts major research centres at risk. The RAE may bring status and charisma, but bad use of its results may erode the quality of the work it identifies and is supposed to promote.

Ron Johnston is professor of geography at the University of Bristol.