A1068
Title: A random-coefficient regression model of co-patenting networks among European cities
Authors: Giovanni Cerulli - CNR-IRCRES (Italy) [presenting]
Antonio Zinilli - National Research Council of Italy (Italy)
Abstract: The aim is to present an application of the responsiveness score model (RSM) to investigate how both network-related attributes (such as degree, closeness, clustering, and betweenness) and contextual factors (including geographical distance, GDP, and GDP per capita) influence co-patenting behavior among European cities. The analysis is based on a rich longitudinal dataset tracking co-patenting networks over the period 2000 to 2020 (source: PATSTAT, RISIS). The RSM employs an iterated random-coefficient regression framework to estimate unit-specific partial derivatives of the number of co-patents with respect to each factor. This allows for capturing the heterogeneous "responsiveness" of each city (node) to different drivers of innovation collaboration. This methodology enables three key contributions: (i) it provides a ranking of cities based on their responsiveness scores, offering a novel lens on territorial innovation dynamics; (ii) it identifies which factors are most influential in shaping co-patenting performance; and (iii) it uncovers the distribution and heterogeneity of responsiveness across the network. The results are accompanied by detailed graphical visualizations that enhance interpretability and facilitate comparative analysis across cities and over time. Empirical findings and their implications for innovation policy and regional development are also presented.
