On the urban bias of patents and the scaling of innovation

While recent studies have heralded large cities as “innovation machines”, the majority of regional studies of innovation are based on patent indicators. In this paper, we compare regional patent and innovation counts in Sweden (1970-2014) and document the presence of a sizeable urban bias in patent indicators, which is primarily explained by higher patent filing propensity in urban areas. We also show that using administrative spatial units which do not account for spatial organization of economic activity tends to exacerbate this bias. This poses a problem for academic studies that wish to understand regional innovation, or policy reports benchmarking regional performance.