Innovation snowballs through path dependence – a wide variety of sectors and data supports the theory

//DH

Nor is path dependency unique to energy innovation. Empirical work, for example, has demonstrated that innovation snowballing affects the direction of pharmaceutical inventions.75 And one carefully conducted study by economist Walker Hanlon provides strong empirical evidence that path dependency shaped the historical evolution of textile technologies.76 In particular, the American Civil War cut off the supply of American cotton, leading demand to temporarily shift to lower-quality Indian cotton. That temporary shock caused a jump in innovations for processing Indian cotton, which, in turn, made the quality of Indian cotton as good or better than American cotton. As a result, the impacts of the temporary innovation shock lasted past the end of the Civil War. Innovations in the technologies for processing Indian cotton continued to increase, alongside demand for Indian cotton, even after American cotton became available again.77

Scholars have also mapped out the innovation stocks of a wide array of different technologies in an analysis of 1.8 million U.S. patents and their citations.78 Consistent with the theory of innovation snowballing, they find that technology classes with more past innovations tend to innovate more.79 Indeed, variation in technologies’ previous patenting activity (i.e., patent rates among “upstream” technologies) explains more than half of the aggregate variation in present patenting activity across different technology classes. This result suggests that the size of a technology’s knowledge stock is highly determinative of that technology’s future innovation trajectory.80