“The role of the good civil servant,” said Sir Humphrey Appleby, “is to find a problem with every solution.”1 In mainstream British life, this appears to translate into a paralysing fascination with complexity.

The greatest of all sources of complexity of current academic life in the UK is undoubtedly the Research Assessment Exercise (RAE), introduced by the Higher Education Funding Councils in 1992. Ratings in the RAE feed directly into the funding formula for higher-education institutions. Run episodically at unpredictable intervals of between 5 and 7 years, the RAE takes into account an eclectic mixture of indicators: papers published (up to four per academic in the 2001 cycle); grants awarded; research students, staff and strategies; and “measures of esteem” (which are not defined on the principal RAE website).

Higher-education institutions in the UK devote millions of pounds’ worth of senior academics’ time to preparing their submissions to the RAE, choosing which members of staff to include in their shop window, and weighing every syllable that goes into a document of a precisely prescribed size. In parallel with this, academe temporarily becomes a giant intellectual meat-market as higher-education institutions vie with each other to buy in staff with impressive CVs in time for them to be included on their RAE return, while experienced individuals whose special skills and vocation lie in university teaching are simultaneously in grave danger of “going under the cleaver”. All this because the key principle of success in the RAE, at least in 2001 (the most recent RAE), was to enter for judgement the maximum possible proportion of academic staff with the highest possible average level of performance in research.

And who are the judges? Panels of one’s peers in individual disciplines are the judges, and their task is to pore over research inputs and outputs that have already been through the peer-review process. Many more hours of academic time are devoted to this rereview of research and to the determination of criteria by which each particular specialty is to judge itself.

Moreover, the rulebook changes each time the RAE is run. In 2001, academics were considered in cognate groups called “units of assessment”, and the maximum possible score was five star (all individuals entered for consideration four published papers of undisputed international quality), as opposed to five (on average, half of the papers entered for consideration were of international quality and the other half were nationally significant). In the end, many higher-education institutions increased their ratings, but the government declined to expand the overall funding cake. Thus, the big fish now consume an even larger share and the little fish are given even less. The rulebook for the next RAE, in 2008, is still being written by a committee chaired by Sir Gareth Roberts, President of Wolfson College, Oxford, but it seems likely that the unit of assessment will be the individual and the highest possible score six stars. Going into the RAE is like entering a football World Cup, but not knowing the number of players, the shape of the ball, the field or the goalposts, or whether there is an offside rule.

Worst of all, university life is dominated by the exercise. While papers of international relevance are required to achieve a five-star rating, work of local relevance, such as most health services or policy research, cannot achieve this standard. Nor can most studies on the impact of new discoveries and implementing new methods in health services or other industries. Methods, rather than their application, are favoured. Particular study designs are deemed creditable in a somewhat arbitrary way, depending on the composition of the discipline panel.

Those of us who work in “research-led” universities are afraid to show too much interest in teaching, as it will not gain RAE points. Thus, the whole direction of university life is affected, with the emphasis on research — or, at least, on certain types of research — rather than teaching. Although this is a research assessment, the implications flow on to teaching, both through the funding available to departments and the interests of people recruited to academic posts. In a recent meeting with the higher-education minister of a Middle Eastern country, one of us was told that this country now sends its students only to university departments with an RAE rating of at least four.

Understanding the rules of the game appears to be as important as the quality of the research. Should your submission favour breadth or depth? Who should be “returned” (ie, put forward by the institution for scrutiny)? Since the result of the process is directly translated into funding, making the right decisions is crucial.

Of course, there are benefits. There is a much smaller possibility for non-productive academics to hide; there is greater transparency about what are the values governing academic success; and there is a real need for individuals, departments, faculties and universities to define their goals and to measure performance. But the price paid for these gains is increasing disgruntlement in the academic workforce and direct lobbying of parliamentarians for a better, less disruptive system to be devised and imposed.

For the record, the Australian equivalent of the RAE, which also affects university funding, combines routinely available data on various kinds of publications with information on grants awarded and PhDs completed. Publications are weighted by “impact factors”, a rating method with many shortcomings, but at least it is in the public realm and known in advance. The system is cheap, because most of the data are already available; it is unselective, because all academics and all of their outputs are considered; and it is simple, transparent and fast — fast enough to be run annually, so that the lag between effort and reward is relatively short, and fast enough to avoid causing a mass moonlight flit of university staff!

Doubtless, some British civil servant found a problem with this “colonial” solution.