Few countries have as much power vested in regional states as Germany. This federalism was a response to the centralization of power and 'Germanization' of education, science and art that occurred under National Socialism in the 1930s. In science and education, as in many other areas, this federalism has created a complicated system in which the responsibilities of the federal government and the 16 Länder (state) governments are intricately interwoven — too complicated to function efficiently. In a bid to revitalize the crisis-ridden country, the German government now plans to simplify matters.

Germany's research minister, Edelgard Bulmahn, may have sensed a political opportunity. Lacking profile in the cabinet, she needs a success. She wants to centralize funding of the Max Planck Society and the DFG, Germany's main research funding agency, while giving the Länder full financial responsibility both for university infrastructure and for non-university research carried out under the umbrella of the Leibniz Association. Her ideas have received a hostile reception from researchers (see page 790 in this issue).

But Germany's research managers are dismissing the idea too hastily. They say that centrally funded basic research could become overly biased towards political and strategic objectives. However, the world is full of successful examples of such centralized structures, in which funding agencies are nevertheless protected from direct political control. Indeed, a streamlining of the federal research system's cumbersome decision-making procedures is long overdue.

Giving the Länder full responsibility for universities, for example, would allow them to find solutions that better match local needs and requirements. This would be a vast improvement over the existing system, in which each new X-ray machine or building project that is backed by the Länder must also be approved at a federal level.

But maintaining and renewing university infrastructure is a costly and thankless task. The federal government's joint responsibility has a regulating effect, keeping in check the disparities in the quality of universities' research. Germany's science system would benefit if the universities were forced to rethink their strengths and weaknesses, and develop more distinctive profiles. The introduction of tuition fees could also support these efforts.

There are pitfalls to be avoided, however. True, the Länder have a vital interest in maintaining universities, just as they have in building and maintaining schools, regional roads and parks. But they would be less enthusiastic about the rest of the legacy. In particular, most Leibniz institutes have little or no regional linkage, so why should the state of Brandenburg, for example, be interested in funding the Leibniz institutes of astrophysics and climate research in its capital, Potsdam? What is to stop it closing these laboratories?

The Leibniz institutes are the least known and most diverse pillar of German research. Bulmahn's proposals are in effect a policy of abandonment that risks eroding Germany's arduously restructured research landscape. Perhaps the best Leibniz laboratories can find a new home at government-funded national research centres, or in the Fraunhofer Society for applied research?

Although more sensible than research managers would like to admit, Bulmahn's proposed reforms have put her out on a limb. In their current form, they might result in a cutback of overall research expenditure, the very opposite of the German government's declared goal. But with more thought and much fine-tuning, the reforms could be good for German research.