Why Don't Inventors Patent?

Abstract:     
A lack of data on innovations without patents has made it impossible to determine whether patenting rates vary systematically across industries and over time. This issue, however, is essential to our understanding of theoretical and empirical analyses of innovation, which typically use patents as a proxy for innovation. This paper takes advantage of a newly-collected data set of more than 8,000 historical innovations to provide the first quantitative estimates of the patenting rates – the share of innovations that are patented. The data indicate that only a fraction of innovations (about 12 to 15 percent) are patented, and that patenting rates vary substantially across industries (from less than 5 to more than 50 percent of all innovations). Patenting rates are lowest in industries where innovations can be easily kept secret and highest in industries where they can be easily reverse-engineered. Such variation suggests that the effectiveness of secrecy, as an alternative to patents, is a key determinant of inventors’ patenting decisions. Comparisons over time confirm this hypothesis: Patenting rates increase in response to scientific breakthroughs that lower the effectiveness of secrecy, both over time and relative to other industries.

Keywords: Innovation, Invention, Patents, Intellectual Property, Economic History

JEL Classifications: O30, O31, O34, N00, L51, L43, K11

Working Paper Series