Modelling the large and dynamically growing bipartite network of German patents and inventors.

The article focuses on analyzing the dynamics and drivers of inventor team formation and innovation using a massive bipartite network of inventors and electrical engineering patents filed in Germany between 1995 and 2015. To manage the large-scale data and account for inventor entry and exit ("actor natality" and "mortality"), the authors propose a Temporal Exponential Random Graph Model (TERGM) with a time-varying actor set and novel sufficient statistics tailored for bipartite temporal networks. Their empirical analysis reveals that factors such as spatial proximity, prior teamwork (team persistence), collaboration interlocking (shared past collaborators), gender, and seniority significantly influence inventor collaboration patterns. Notably, while male and female inventors show similar propensities to invent, there is weak evidence of gender homophily among female inventors. The study contributes both methodologically and substantively by simultaneously modeling multiple mechanisms of inventor collaboration in a comprehensive population dataset, offering insights relevant to innovation research and policy.