Abstract: Using the methods of exploratory spatial data analysis(ESDA), this paper analyzes the spatial distribution and dynamic changes of patent outputs in China, by studying 341 prefecture-level cities from 1997 to 2009. Then, the paper examines the spatial club convergence hypothesis of patent outputs by using spatial econometrics model, and finally explores the formation mechanism. The ESDA analysis using the Moran's I gives the strong evidence of spatial autocorrelation of patent outputs among cities, and the aggregation level has increased steadily over the past years. However, in the H-H and L-L groups, with the growth and the degree of aggregation increase, the difference of patent outputs between the two groups continues to expand and then form two "spatial convergence clubs". The empirical results show that the spatial heterogeneous and local geographic spillovers factors contribute to the features of overall differentiation and local multi-steady-state of regional patent outputs in China. Finally, this study presents some references valuable for analyzing spatial structure and policy-making.

Key words: patent outputs, ESDA, spatial dependence, geographic spillovers, spatial club convergence