Does it matter whether research is conducted by the private business rather than in universities or government research centres? While most of the attention of science and innovation policy in the last decades has explored the relevance of the interconnections between public and business players in enhancing knowledge-based societies, a major trend has been ignored: both the quota of public R&D and its share over the total R&D investment has shrunk in the majority of OECD countries. As a result, a larger fraction of knowledge is today generated in the business sector. We argue that this is a major problem since public research and private research differ along a number of characteristics, e.g. public access, potential for future technological innovations, criteria of resource allocation. This trend can have adverse implications for long-term innovation and economic welfare in our societies. Through the lens of the public goods theory and of the sector of funding and execution of R&D for the period 1981–2013 we try to explain why.