Sir

You reported the finding that the scientific performance of research institutes supported by the UK Biotechnology and Biological Sciences Research Council during 1981-94 was higher than that of a group of 15 British universities (Nature 390, 12; 1997). But a comparison between the output of universities and other research organizations does not necessarily provide much insight into the factors responsible for observed differences.

We have carried out a study of the effects on research performance of changes in the funding structure of Flemish universities in the 1980s and early 1990s (available on request). In Flanders, as in many countries, universities receive a ‘basic allowance’ which they are relatively free to spend as they choose. Since the end of the 1970s, this allowance has remained virtually constant.

At the same time, the government has made additional funds available to university research groups on a competitive basis. You have also reported proposals by the Dutch Minister of Education, Culture and Science to transfer DFL500 million (US$250 million) of university funding to the National Research Council (Nature 390, 9; 1997).

We established for 345 research departments the amount of research funded from the basic allowance, and the support from external funds, expressed in Full Time Equivalents spent on research (FTE-research). We collected bibliometric data and examined trends in publication productivity, defined as the number of articles per FTE-research, as well as citation rates.

We found that, during the 1980s and early 1990s, externally funded research increased by 7 per cent a year. It became increasingly concentrated in a small number of departments, while the distribution between departments of research funded from the basic allowance remained constant. In 1989, 18 departments — 5% of the total — accounted for 37% of the externally funded research capacity, but only 11% of the ‘basic’ research capacity.

Departments with a high international impact have profited much more from external funding than groups with a lower standing. The general aim in Flanders, to introduce more competition into the allocation of funds for academic research, has been successful. But overall productivity has remained constant, mainly because the publication output of the departments with the strongest increase in externally funded research decreased significantly, whereas the ratio of junior to senior scientists increased dramatically, from 1.6 to 3.9.

If these trends continue, a situation may emerge in Flanders in which the research base, especially tenured personnel, normally provided by a university out of its own resources, becomes too small for externally funded research activities. Furthermore, there appears to be a limit to the extent to which control over research funds can be transferred from universities to external funding agencies without a loss of productivity. If this process is pushed too far, the productivity of the system as a whole may decline, as the conditions in which senior scientists have to work — particularly in departments of high international standing — may make it impossible to maintain the potential quality of their research or to train young scientists.

Also, the increasing importance of external competitive funding may lead to an imbalance in the funding system unless universities develop their own somewhat risky internal research policies, relying more than normally upon external funding patterns. This is illustrated by the fact that the faculty of medicine at the Catholic University of Leuven, which has a strong internal research policy, showed the highest impact and productivity of all faculties in the study.