The aim of this paper is to analyze the spatial distribution of innovation, measured by patents per capita, in geographical micro-regions of Brazil. This purpose is achieved by using exploratory spatial data analysis. This paper is pioneer in adopting this set of statistical tools when we look at the Brazilian innovation literature. Our findings reveal that the patents per capita are distributed non-randomly, suggesting positive spatial autocorrelation. The spatial heterogeneity confirms the existence of North-South polarization regime in terms of technological activity in Brazil. Local indicators of spatial association are also computed as well as bivariate Moran statistics as means of assessing the influence of the neighborhood on the technological performance of each microregion. We find out that the lack of some elements such as industrialization, higher-educated population, PhD researchers, workers with technological background, and technological and industrial diversity in a given microregion may be working as spatial barriers, thus precluding the diffusion process of knowledge spillovers.
exploratory spatial data analysis; patents per capita; spatial autocorrelation; spatial regimes; Brazil
