A "radical overhaul" of research assessment that will decide the fate of £1 billion of public funding will be placed before the funding councils next week.
It consists of series of recommendations rather than a single alternative model. It proposes that research directors be tipped off about their chances of success before deciding whether to have research peer reviewed.
Universities would be ranked according to measures such as number of publications, external research income and number of PhD students in each unit of assessment.
Individual institutions would decide whether to accept a basic research grant based on these metrics or to apply for a higher level of funding and risk losing the lower grant.
The proposal has emerged from the review of research assessment conducted by Sir Gareth Roberts, president of Wolfson College, Oxford, on behalf of the funding councils.
He told The THES : "The strongest message to emerge from the process has been the indispensability of expert peer review to the assessment of research. There has been a general acceptance that the burden this imposes is both necessary and unavoidable. The report will recommend a number of major reforms to the research assessment exercise, which, taken together, represent a radical overhaul of the system."
The report will consist of series of separate recommendations, some of which will be viable only if other reforms are approved. The first assessment would be in 2008. It would be repeated every six years.
The report recommends that, if a department or faculty were to be peer reviewed, all staff should be assessed. This would avoid excluding young researchers. Assessment panels would decide how many staff in such departments were internationally excellent and should be funded.
It is anticipated that a third of units will be funded through the "research quality assessment" route, with the rest getting cash through the "research capacity assessment" route. The RQA route would attract five times as much money as the RCA route.
The draft report will be considered by the boards of the funding councils next month. Formal proposals will be published for consultation in late May to September. The funding councils will decide in autumn whether to adopt them.
It is hoped that the RQA will continue to be identical in England, Scotland, Wales and Northern Ireland, even though individual funding bodies have chosen to fund the results differently.
* Education secretary Charles Clarke this week scotched rumours that the government intends to abolish funding for departments rated 4 in the 2001 RAE.
He told the Commons education select committee: "It's not true that we want to cut out all the funding (for 4-rated departments), but we want to assure ourselves that 3a and 4s have aspirations to world-class research."