This study investigates the efficacy of public R&D support. Compared to most existing studies, we do not stop at substitution effects or general innovation outcome measures, but we are interested in knowing where the policy effect is highest: on innovation close to the market (i.e. incremental innovation) or on innovation that is still far from the market and hence more risky and radical. Using firm level data from the period 1999 to 2011, we find that the policy hits where the market failure is highest, that is, for radical innovation. Taking into account that the Swiss funding agency encourages collaboration, we find no evidence that the impact of the policy is positively effected by various R&D collaboration patterns