Forward patent citations as predictive measures for diffusion of emerging technologies

Abstract

An extensive literature traces diffusion of knowledge and inventions using backward patent citations - i.e. patents cited by a new patent. Backward citations show how knowledge imbedded in past inventions influences new inventions. What can we learn from forward citations? Can forward citations be used to forecast the evolution of emerging technologies? In this study we analyze forward citations for three classes of technology - biotechnology, telecommunications and alternative energy. We use patent metrics to measure the rate of technological innovation. Following the existing literature, we examine the incremental advances in the three selected technological fields as indicated by forward patent citations. Our initial proposition is that forward citations should follow a classic S-curve distribution. The paper examines this proposition. Our study entails: selecting groups of patents issued at least 20 years, and some as far as 30 years ago; extracting the number of forward citations issued each year for each patent through 2008; and fitting the cumulative forward citations to a variety of models such S-, linear, quadratic and logistic curves. The analysis and findings are discussed in the paper.