The determinants of parallel invention: Measuring the role of information sharing and personal interaction between inventors

Historical accounts describe numerous cases of parallel invention. Nowadays, with over half a million inventions yearly that apply for patent protection at the USPTO alone, it is likely that there are a lot of parallel inventions among these. Yet, the mechanisms behind creating similar knowledge remain unstudied. From both a theoretical and practical perspective, it is an interesting question to what degree parallel inventions take place truly independent of each other, or whether they are the result of the exchange of knowledge and ideas between inventors. In our empirical study, we use the unique setting of technical standardization, where it is possible to systematically observe knowledge sharing as well as knowledge exchanges between inventors in detail. This study presents two novel analyses, one focussing on the determinants of similar inventions (using an AI-based approach) and one on the determinants of identical inventions (exploiting data from the patent granting procedure). In both analyses, we find positive and significant effects for knowledge sharing as well as for inventor interaction as determinants. The latter effect is the strongest: if meet in person and discuss their ideas, the likelihood of similar inventions increases up to a factor of approximately five, to up to 2.3 percentage points. Empirically confirming the theoretical work of Amabile (1983, 1988) on knowledge creation at the individual level and that of Nonaka (1994, 2006) on knowledge creation at the organizational level, we reflect on the implications of our findings for companies wishing to increase their inventive efforts.