Observing Technological Trajectories in Patent Data: Empirical Methods to Study the Emergence and Growth of New Technologies

Abstract

finement of new commercial processes and products. The concept of a natural technological is a synthesis of heuristic evolutionary economic notions, generalized from numerous casestudy observations across many industries, that the ebbs and flows of technological change tend to follow particular pathways along which progress moves more quickly. Many of the interesting economic relationships that drive technological change in industries are influenced by the sequential emergence of incremental technological refinements over time and by the interaction between different currents of technological advance. The need to account for the complex and cumulative nature of separately evolving technologies in the agricultural biotechnology industry has led to the consideration of empirical methods that use patent data to identify and track the evolution of separate technological trajectories. Patent documents provide some of the best data available on R&D output (Griliches). However, a primary drawback of using patents to track the micropatterns of innovation is, of course, that not all new ideas are patented, especially given the differences in institutional significance of patents across the different sectors of the R&D economy, including universities, startup companies, and corporations. The examples that follow utilize a dataset of U.S. patents on biological inventions with relevance for crop agriculture. These methods to empirically observe technological trajectories have proven valuable for three reasons. First, the methods provide for a well-selected dataset, specific to the research question at hand, and minimizing, as much as possible, problems of selection bias, missi g data, and truncation. Second, these methods can allow for structural modeling of causal relationship between innovations. Third, the methods provide for the theoretically correct control of technology field effects in regression an lysis: all of the individual innovations contained in a single trajectory sample can be expected to arise from a common technological opportunity set, to have common intrinsic appropriability characteristics and face common external appropriability conditions, to face the same regulatory environment, and to face similar market demand.