Education: From him that hath not...

If you want to do research, go to a research institute. The teaching and bureaucracy get in the way in universities." So I was told when I applied for a lectureship in 1981. Then the cuts in cash started, and student numbers doubled - twice. My first class contained some 50 students. By 1996 the course had 240, and half the staff had taken early retirement.

During the Eighties, as student load increased, many academics came to agree that research institutes were nicer places to be. Simultaneously the government made the polytechnics into universities. However, the Treasury did not want to fund the ex-polytechnics at the level of the old universities. This coupled well with the desire of many academics to be in well funded research institutes with minimal teaching. Academics with influence wanted all the research money and few undergraduates, and the Treasury wanted as little research expenditure as possible. Thus the research assessment exercise (RAE) was born, to satisfy both parties.

And so, in science at least, the RAE is all about money. It follows Sir Humphrey's dictum (from Yes Minister) that you can judge a department's success by how much it spends. The keys to success in science are apparently money, research students and publications, but this measures money three times. Money pays for researchers, including students, who produce data for publication. Any fool with money looks good, and among the genuinely excellent researchers there are some upmarket used-car salesmen.

The "eminent" scientists who developed the RAE criteria typically came from rich departments, even research institutes; which goes a long way to explaining their eminence. This is like letting Manchester United and Liverpool decide how to allocate football's TV revenue throughout the league clubs. The stated intention of some is that there should be only 10 first-division universities doing research, just as in the good old Twenties, so the RAE gives more money to those who spend most.

If the RAE wanted to be objective, it could look for "value added" and divide publications output by expenditure input. That would make expensive grants a responsibility, and save money; it might also make some of the big-spending universities look less successful. A comparison of a 3a department's publications with a grade 5s, using the wonders of IT, showed surprisingly little difference in output, but the grade 5 spent four times as much. Try IT if you doubt me. A colleague whose department went from 4 to 5 last time attributed this to an extra pounds 2m EU grant rather than to improved quality.

The RAE could also divide the value of each publication by the number of authors, and insist that each author gets her due credit. This would reduce parasitism by senior staff on subordinates.

The premise that the RAE continually increases research performance is flawed, in two ways at least. First, the cash pot is fixed, except for EU money and industrial investment. Every RAE point gained by one university is money taken from all the rest. The taxpayer gets no more research, probably less. Taking a little from an underfunded researcher may stop him altogether, while it is an insignificant addition to a rich one's budget. Also, we all spend more on playing the RAE game.

University staff need to write more than five grant applications a year each, and swamp the awarding bodies. Apparently it is not worth trying to do a good job. Rejected applications, given a few cosmetic changes, may be funded if you throw a six. In one case just the name of the enzyme was changed before successfully resubmitting to the same research council. Advice from a committee member: get to know committee members, find out what they like and submit requests in those areas, but don't compete with them.

The second flaw is a basic Thatcherite misunderstanding of Darwin's theory of evolution. If you have a savanna supporting 10 lions, and you cull the five weakest, the rest breed, and you still have 10 fit lions. If you have 10 institutions, public or private, and you cull the five weakest, you get five bigger institutions and less competition. They do not reproduce; they just grow; and if a baby is born, they eat it. Ask Freddy Laker.

Meanwhile the RAE is a compulsory game, with rules to be exploited. Choice of assessment unit can be critical. Putting an ordinary biochemistry department into "subjects allied to medicine" could be a spectacular coup. Ordinary universities must decide whether to go for 80 per cent of staff and a 3, 65 per cent of staff for a 4, or one person for a prestigious 5. (Yes, there is a 5-rated, one-person department.) Now, of course, the rich departments with few undergraduates want all staff to be entered, to increase their own standing, and are opposed to this.

You can survive only if you have a "patron" - industry, a grant-awarding body, or an eminent academic. Departments have clones of researchers doing the same thing, and large parts of the undergraduate syllabus have no specialist teachers. Researchers are avoiding teaching, and teachers are being "invited to leave". Departments with too few postgraduate students to provide teaching assistance have classes containing hundreds of undergraduates. But the certain result of giving more to those who have most is to turn a few universities into research institutes, and most of the universities back into polytechnic student-factories.

The writer is a lecturer in biology and biochemistry at Queen's University, Belfast