Territorial Patterns of European Innovation in the Context of Different Innovation Output Proxies: A Spatial MGWR-SAR Approach

Abstract The paper seeks to explore the drivers of European innovation represented by three innovation outputs (patent, trademark, and design applications), emphasizing spatial autocorrelation and heterogeneity. It includes data from 202 regions from 22 European Union (EU) Member States, along with 18 regions from Switzerland, Norway, and Serbia in 2019, providing a more comprehensive geographic scope. By considering multiple indicators of innovation output, including patents, trademarks, and design applications, the main objective is to examine spatial innovation spillovers and the heterogeneous responses of regional innovation output to innovation inputs in the context of European regions. To achieve this goal, the main instrument of the analysis is a newly proposed methodological framework called Mixed Geographically Weighted Regression-Spatial Autoregressive (MGWR-SAR) models. The analysis suggests that while all innovation inputs (most-cited publications, research and development expenditure in the business sector, human resources in science and technology, and population density) are justified in increasing all innovation outputs, the strength of particular determinants of innovation might vary across regions. Moreover, the analysis reveals valuable insights into how spatial spillovers influence regional innovation. The impact of spatial connections varies across the regions, with patents showing the strongest linkages, affecting about 92.27% of regions. Although trademarks and designs have fewer spatial connections (approximately 50% of regions), they still play a significant role in innovation. Although patents have traditionally dominated discussions of innovation, the findings reveal the importance of incorporating designs and trademarks as complementary indicators. Overall, the study highlights the need for multiple metrics to comprehensively evaluate innovations.