Novel science for industry?

Measuring novel science as publications which make new combinations of referenced journals and measuring links between science and technology by scientific references in patent applications, we explore the complex relationship between scientific novelty and technology impact. We draw on all the Thomson Reuters Web of Science journal articles published in 2001 and all the patents in PATSTAT version 201310. We find that only a small proportion (about 10%) of all scientific publications are referenced as prior art in subsequent technological inventions, but a small number of scientific papers which score on novelty (about 11%) are significantly more likely to have technological impact, particularly the 1% highly novel scientific papers. The technological impact premium for novel scientific papers is even bigger, when correcting for their initial disadvantage of being less likely published in high impact factor journals. In addition to this superior direct effect, novel science also has a higher indirect technological impact, being more likely to be cited by other scientific papers which have technological impact. Within the set of scientific papers cited at least once by patents, there are no additional significant differences in the speed or the intensity of the technological impact between novel and non-novel scientific prior art, but the technological impact from novel science is significantly broader, covering more diverse technological fields and reaching technology fields previously non-impacted.