Japanese patent network analysis reveals tech sophistication fuels regional growth

Researchers Rintaro Karashima and Hiroyasu Inoue analyzed 3.9 million corporate patent records from 47 Japanese prefectures across 35 technology classes (1981–2015). Using the Fitness-Complexity algorithm—originally developed to measure national economic complexity—they derived regional 'Fitness' scores for seven five-year periods. Their fixed-effects panel model with Driscoll-Kraay standard errors shows a significant positive association between prefectural Fitness and subsequent five-year real GRP per capita growth (β = 0.0029, p = 0.007), controlling for initial income, population density, and overall patenting activity. Notably, this relationship only emerges when both entity and time fixed effects are included, highlighting the importance of controlling for unobserved regional heterogeneity. Cross-sectional correlations actually reverse sign across periods, underscoring the panel approach's necessity.

Crucially, the growth effect of Fitness is stronger in prefectures with lower initial income, implying that catch-up regions benefit disproportionately from building sophisticated technological capabilities. Lag and lead analyses confirm that the direction runs from Fitness to subsequent growth, not the reverse. This suggests that targeted investments in complex, diverse technologies—beyond simple patent counts—can accelerate economic development, especially in less advanced areas. For tech policymakers, the implication is clear: fostering a rich portfolio of interconnected technological skills may yield outsized growth dividends, particularly in regions with room to expand.