Patents, technological inputs and spillovers among regions

Abstract

This article analyses the importance of different technological inputs (R&D and human capital) and different spillovers in explaining the differences in patenting among Spanish regions in the period 1986 to 2003. The analysis is based on the estimation of a knowledge production function. A region's own R&D activities and human capital are observed to have a positive significant effect on innovation output, measured by the number of patents. R&D spillovers weighted by the distance and the volume of trade flows between regions cause positive effects on a region's patents. However, distance matters more than the intensity of trade flows and the R&D spillover effects between regions are bounded: spillovers from closer regions perform better than spillovers from distant regions. On the opposite side, human capital spillovers do not cause any effect outside the region itself.

Notes

¹Coe et al. (1997) use import shares of intermediate goods as weights, while in Coe and Helpman (1995), R&D spillovers are constructed as a weighted average of the domestic R&D of trading partners using bilateral import shares as weights.

²See, for example, Engelbretch (2002). Using human capital stock variables, he confirms the existence of a positive role for human capital in the absorption of international knowledge spillovers other than embodied R&D spillovers.

³As proxy variables for human capital, Rondé and Husller (2005) use the percentage of population devoted to R&D, and Bottazzi and Peri (2003) use the share of college graduates in the population.

⁴The variance of the negative binomial model is equal to the variance of the Poisson model multiplied by the coefficient of over-dispersion, Delta. The limit case where Delta equals zero corresponds to the Poisson model. Hence, to test the appropriateness of the negative binomial regression, we provide the Delta statistics in the tables of results.

⁵Rondé and Hussler (2005) test the influence of the regional stock of knowledge on regional innovative dynamism (patents) using different indicators of human capital (percentage of the population devoted to research). Bottazzi and Peri (2003) also analyse the importance of human capital (proxied by the share of college graduate in the population of a region). Their results show that the effect of human capital on innovative output is significant.

⁶Results are available upon request to the authors.

⁷Unfortunately, the available statistical information for the Spanish regions does not allow us to analyse the effect of other regional characteristics on the innovative activity.