The evolving boundary of green technology

Abstract

Green patent indicators are widely used to track technological progress, assess climate and innovation policies, and identify emerging technological opportunities. Yet these indicators depend on classification systems that evolve over time and shape which inventions become visible. Drawing on a sociological perspective, we interpret patent classifications as information infrastructures and examine three mechanisms that shape the visibility of green inventions: (1) modification of category boundaries through reclassification, (2) broadening of institutional coverage through set expansion, and (3) narrowing of scope through user-defined selection criteria. Using patent families filed between 1980 and 2016, we merge the 2019 and 2023 PATSTAT releases to isolate their effects. Reclassification and set expansion increase the observed population of green patents by about 9% and 10%, respectively, while commonly used filters exclude between 42% and 93%. These effects are uneven and particularly pronounced for recent inventions, rapidly growing fields, and patents associated with Asian offices. Patents made newly visible through reclassification and set expansion are more novel and more often cited, while being slightly more specialized in narrow technological fields. Measurement choices can therefore change conclusions about the pace, direction, maturity, geography, and nature of green innovation. Observed green patent trends reflect not only inventive activity but also the evolving infrastructure through which green technology is made visible. Evaluations of policy effects, technological leadership, emerging opportunities, and progress in the green transition are therefore sensitive to classification vintages, institutional coverage, and filtering choices.