We show that the presumed incompatibility of uncoordinated R&D and competition is not fundamental, but hinges on the nature of R&D spillovers. As a consequence, R&D subsidies may be more effective than previously thought.