Whatever happens over the next few months in budget negotiations for the European Union's Seventh Framework Programme for research, there now seems to be widespread political acceptance that the European Research Council (ERC) should be established as a funding body for the continent's researchers in the sciences and humanities.

The most visible step so far in the creation of the ERC occurred last week when the 22 members of its scientific board were named (see http://www.europa.eu.int/comm/research/press/2005/pr1807en.cfm). The European Commission's selection committee for the board did a good job, combining strong scientific credibility with a reasonable geographical balance. The commission now needs to build on this success in its creation of a well-crafted agency.

Contrary to the wishes of some member states, the ERC's foot soldiers should not primarily be seconded from national agencies; they should be independently recruited. To keep the disciplinary programmes vital and responsive, they should be led by scientists from research institutions. And above the heads of the ERC executive staff, the agency should be placed firmly in the hands of the commission during its first few years, rather than be controlled more directly by member states.

Such a structure will be not unlike that of the US National Science Foundation (NSF). It is not a perfect system, but it has served US science well for more than 50 years and it has the sort of autonomy to which the ERC should aspire. All would agree that, like the NSF, the ERC should be a competitive organization for bottom-up basic research, funding researchers from across all disciplines and from across the continent, solely on the basis of excellence.

Experience demonstrates time and again that when Europe’s member nations become involved, their individual goals distort agendas and delay decisions.

But ‘autonomy’ means different things to different people. For some, it means as complete a dissociation as is possible from the European Commission. History might seem to support this argument, even though it will be the commission that will be footing the bill, likely to be up to €1.5 billion (US$1.8 billion) per year. The commission's customary bureaucracy has never failed to alienate researchers.

But the commission can in this case play a useful role in sheltering the new-born agency from interference by competing national interests. Opponents of management by the commission have latched on to a hitherto unused article in the rules governing commission research, which allows a separate agency to be set up to run a particular programme. At an informal meeting of research ministers of the European Union's 25 member states in Cardiff earlier this month, opinion was equally divided between the merits of this and a commission-led agency.

It may already be too late for the more independent agency to be in place at the start, if the ERC is to be launched in 2007 as planned. Its establishment and location would require approval by both the European council of ministers and the European parliament — bodies that stretched out the decision to create the somewhat analogous European food agency over three years in an acrimonious and heavily politicized struggle.

Even more dangerously, there is a strong possibility that the agency would establish a decision-making board filled with national representatives. In contrast, a commission-run agency would have only a five-strong steering committee to ensure that it executes appropriately the will of its scientific council.

European experience demonstrates time and again that when the European Union's member nations get too closely involved, their diverse individual goals distort agendas and delay decisions. One of the commission's cardinal roles is to shelter the implementation of European policies from the distorting effects of national interests.

The best way forward is therefore to set up the ERC as a commission-led agency with the option of switching to a more independent operation later if it proves desirable. The commission would then have time to prove that it can run an agency capable of executing the desires of its scientific council without debilitating bureaucracy. Better to live in hope than to sacrifice the ERC to the certain buffeting between the conflicting fancies of Europe's 25 member states.