In The THES (November 14), Sir Robert May, Bahram Bekhradnia, "John Smith" and others reflected on the future funding of research by the ÖйúAƬ funding councils. Bekhradnia suggested the choice was between selective funding for research or "spreading the money thinly, without regard to quality". That choice is no choice.

Research assessment exercises have put an end to the unworthy academics who could be found a few decades ago. However, they have also increased stress among academic staff. After all, we have seen academic staff leave on "voluntary terms" and entire departments closed after RAEs.

Perhaps we can accept that past RAEs have been useful. High-quality research is one of the main expectations of academic staff, and RAEs have shown where this is being achieved. Now the funding councils must consider what should be their roles in research. The vision I propose has two aspects. First, to ensure that their responsibilities within the dual support system are discharged. Second, to fund the development of emerging subjects and departments. Within the dual support system, the funding councils' responsibilities are to provide overheads linked to research grants and to provide the "well-found laboratory".

In their grants and awards, research councils meet neither the salary of the principal investigator nor general overheads. These costs should be determined by discussion between the funding councils, the Committee of Vice Chancellors and Principals and the research councils. Funds could then be provided based on success in gaining peer-reviewed, project-based income, without an RAE.

ÖйúAƬ

Today there are sources of peer-reviewed, project-based income to universities other than research councils, such as the European Union and the medical charities. These provide lower levels of overheads than the research councils. The funding councils should agree what are their dual-support responsibilities with respect to these sources.

As the United Kingdom is already successful in gaining EU funding and the levels of charity and research council support are unlikely to rise hugely, this proposal would provide a means of transferring research funds between universities on a clearly understood basis.

ÖйúAƬ

The great need for capital equipment, clearly identified in the Survey of Research Equipment in UK Universities (June 1996), is not well served by an RAE. Almost by definition, the demand for quality-related funding that follows a successful RAE submission is quite local within a university. The university may need new capital equipment, but this usually comes second to the need to employ more researchers. Further, an annual quality-related allocation at departmental level may be insufficient to make major capital purchases.

The use of existing funds will have to be rebalanced. The annual capital requirement can be derived at the national level: the total requirement identified in the surveyJis about Pounds 470 million.

The source of such funds, perhaps at the rate of Pounds 50 to Pounds 80 million a year, is likely to be contentious, but a solution must be found. The research councils have funded big national facilities, equipment pools and some capital equipment. Funding council support for the "well found laboratory" is less than it should be. Once the HEFCs and research councils decide how much each should provide, panels can be set up to allocate the funds Here, big is best, but with a regional accent. The funds should go to departments that consistently attract research grants and contracts. But regional considerations should be taken into account, and an element of wider access to major equipment must be built in.

The HEFCE should also fund the development of research in emerging subjects and departments. This function is not "research assessment" but rather "research potential assessment". This is research not sufficiently developed to be ready for peer assessment or industry funding. The aim would be to develop novel programmes and departments to the point where submissions for funding can be made with the necessary experience and performance needed for their peer assessment.

ÖйúAƬ

To be eligible for research potential assessment (RPA), a university would have to make a case for its research potential and its strategic thinking. It would have to demonstrate that its research potential is worthy of support, describe the developing subjects, and show that such support would be well used. Departments getting quality-related funding would not normally be eligible for RPA.

I suggest that there should be no more RAEs. This would be a relief to most academic staff: they could get on with teaching and research knowing that the latter would be fully funded. Policy-makers would have to resolve some knotty questions, but these need addressing anyway. The RPAs would be on a much smaller scale than RAEs. Bearing in mind the long time-scale for research development, they would need to be on a longer cycle, perhaps every six years. The overall savings in university staff effort from these proposals might, on their own, enable more research to be done.

John McGinnety is a former pro vice chancellor of the University of East London.