Measuring follow-on innovation

How patents affect follow-on innovation is a key question for the patent system. We disaggregate follow-on innovation into activities that infringe patents and others that do not infringe but can be indirectly affected by patents. Replicating an important study using our disaggregated measure, we find that 87 percent of follow-on scientific publications describing patented genes do not constitute patent infringement. Supplementing our empirical strategy with data on patent expiration dates, we find that gene patents which are not close to expiration cause an increase in noninfringing follow-on research, but the effect disappears for patents close to expiration. Our nuanced measure helps better identify the mechanisms of patents' effect, reconcile disparate results in the literature, and evaluate policy reform.