Abstract Economic analysis has helped us understand the strong economic dimension in the explosive growth of science, and (more recently) the reasons for continuing public subsidies. However, the growing domination of the “market failure” approach has led to the analytical neglect of two major questions for policy-makers. How does science contribute to technology? Are the technological benefits from science increasingly becoming international? On the former, too much attention has been devoted to the relatively narrow range of scientific fields producing knowledge with direct technological applications, and too little to the much broader range of fields, the skills of which contribute to most technologies. On the latter, national systems of science and of technology remain closely coupled in most major countries, in spite of the technological activities of large multinational firms. Empirical research is needed on concentration, scale and efficiency in the performance of basic research, where techniques and insights from the applied economics of industrial R&D are of considerable relevance. There is no convincing evidence so far of unexploited economies of scale in basic research. This evidence shows that many policies for greater “selectivity and concentration” in basic research have been misconceived. Economists and other social scientists could help by formulating more persuasive justifications for public subsidy for basic research, and by making more realistic assumptions about the nature of science and technology.