ÖйúAƬ

Even if John Randall is right in arguing that differential fees would require more intensive inspection and evaluation ("QAA takes tough line on top-ups", THES, January 12), he cannot seriously claim that the Quality Assurance Agency could deliver this.

Already despite its current less intensive arrangements the information the QAA produces is on average four years out of date. The agency has too few subject reviewers to be certain of maintaining even this. Nor has it managed to devise an assessment methodology stable enough to allow comparisons over time.

This overstretch is at least in part the result of the QAA's inability to distinguish between regulation, regulatory enforcement and evaluation. The Financial Services Authority, to which Randall alludes, does not evaluate, it merely produces regulations and codes of practice that guard against conflicts of interest and misleading representations. That is probably a model for us too.

We could get what is needed for any fee regime, cost-effectively and up to date, by making the ÖйúAƬ Statistics Agency collect more detailed statistics and publish them freely. The QAA could retain an FSA-like role: institutions would have to comply with its regulations and codes of practice, but non-compliance would be dealt with through the courts. Evaluation should then be left to private providers (as in the financial sector) that are subject to the market discipline of finding out what their clients want and are prepared to pay for.

The way in which the QAA has implemented the Dearing requirement for clear descriptions of degree programmes - by demanding specifications so abstract and full of jargon as to be useless to parents, students and employers alike - shows that it is institutionally incapable of delivering evaluation cost-effectively.

Tim Reuter Professor of history University of Southampton