In our work on Science 2.0, one of the question is:
Is science 2.0 or open science more effective?
One paper starts discussing the issue by comparing the open science and the traditional approach, hinting about similar results being achieved with less time and money.
However, the evidence is thin.
Are there more robust studies on this?
Otherwise: one idea would be to launch a counterfactual research-action.
For every research project funded by FP7, reserve 20K Euros for a competition on an open innovation platform (e.g. innocentive, sci-starter) to solve the same problem.
Or for a subset of project, invest the same funding in an open science way.
Then compare the results of the two approaches.