An article in Nature reports on a study by Ulf Sandstrom and Peter van den Besselaar on Funding, evaluation, and the performance of national research systems in the Journal of Infometrics. In their study, Sandstrom and Van Den Besselaar have looked at the added value of additional funding to the production of top-quality research (in terms of field weighted citations). One of the reasons for focusing on change was that the comparison of input levels between countries is problematic. For example, in the UK PhD students are not counted as costs (=inputs) significantly improving the input/output ration compared to countries that do count PhDs as costs. Looking at the impact of changing inputs on changing output solves that problem.
The researchers found that where traditional input/output studies show the UK system as the most efficient one and the Netherlands system much weaker, the analysis of change actually shows the UK and e.g. Israel at the bottom end and the Netherlands and Belgium at the top.
Their findings indicate that efficient science systems have a well-developed ex-post evaluation system combined with considerably high institutional funding and not too large university autonomy. Whether autonomy from the state has a positive effect remains unsettled. Countries with an ex-post evaluation system are more efficient, but much less so when ‘strong’ evaluation systems have direct funding effects.
Less efficient systems seem to have strong ex-ante control, either through a high level of so-called competitive project funding and/or through strong power of university management.