A novel technology-industry concordance table based on linked inventor-establishment data

Mapping technologies into industries is frequently required in empirical innovation studies, but many concordances only provide coarse mappings. We develop a novel concordance between industries and technologies making use of linked inventor-employee data for a large sample of German patents filed at the European Patent Office between 1999 and 2011. Inventors listed on these patents are matched and disambiguated with German social security records. Employment data recorded in this database include detailed industry codes describing the industrial activities of the inventors' establishments. These linked microdata allow us to identify the precise industrial origin of inventions, combine them with technology classifications from the inventors' patents and to generate novel concordance tables. We evaluate our approach by comparing the concordances with existing work. Moreover, in an empirical application, we find that patent based indicators represent valid proxies for the innovation performance of industries and thus help to measure innovation in the absence of other survey based indicators. Service sector innovation activities, however, appear to be barely represented by the patenting activities of inventors and establishments.