Spatial Distributions and Determinants of Regional Innovation in China: Evidence from Chinese Metropolitan Data

This article, using a panel dataset covering patents granted in 336 cities and the economic and employment data of 282 cities in China, presents an preliminary exploratory spatial data analysis by Gini coefficient and Moran’s I analysis, and a confirmatory spatial data analysis by spatial Durbin model. We first investigate China regional innovative activities by three different types of patents at metropolitan-level data and make several key findings. First, the spatial autocorrelation of invention patent is insignificant from 2001 to 2013, whereas the coefficients of spatial autocorrelation of utility patents and design patents are continuingly rising across years. Second, the innovation clusters are vanishing in China’s western and northeastern cities, whereas booming in the periphery cities of Shanghai and Guangzhou during 2000–2015. Third, the cities surrounded by high level of GDP output and R&D expenditure neighbors will more likely perform better in innovative activities. By employing smaller territorial units, we provide more specific details about the regional distribution and the dynamic interaction of innovative activities across cities in China.