This paper estimates the effect of research externalities across space in generating innovation using R&D and patent data for 86 European regions from 1977-1995. The results show that spillovers exist for regions within 300 km of each other, but are small in magnitude. Doubling R&D spending in a region increases innovation in other regions within 300 km by only 2-3%, while increasing the region's own innovation by 80-90%. These spillovers are interpreted as the result of local diffusion of non-codified knowledge, embodied in people and spreading via personal contacts. The findings suggest that spillovers are weaker across national borders. The study uses patent counts as a measure of innovative output and finds that R&D spillovers are identified as the effect of R&D intensity in one region on R&D productivity in another. The results show that R&D spillovers are small and localized, with most benefits accruing to the region that spends R&D resources. The study also finds that R&D spillovers are more significant for regions that are closer to each other. The paper also considers the role of human capital in innovation and finds that higher levels of human capital are associated with higher levels of innovation. The study also examines the impact of distance on R&D spillovers and finds that the effect of R&D on innovation decreases with distance. The results suggest that R&D spillovers are strongest for regions within 300 km of each other. The study also considers the impact of historical population on R&D spillovers and finds that historical population is a good instrument for estimating R&D spillovers. The results suggest that R&D spillovers are more significant for regions with higher historical population. The study also finds that R&D spillovers are more significant for regions that are closer to each other. The results suggest that R&D spillovers are strongest for regions within 300 km of each other.This paper estimates the effect of research externalities across space in generating innovation using R&D and patent data for 86 European regions from 1977-1995. The results show that spillovers exist for regions within 300 km of each other, but are small in magnitude. Doubling R&D spending in a region increases innovation in other regions within 300 km by only 2-3%, while increasing the region's own innovation by 80-90%. These spillovers are interpreted as the result of local diffusion of non-codified knowledge, embodied in people and spreading via personal contacts. The findings suggest that spillovers are weaker across national borders. The study uses patent counts as a measure of innovative output and finds that R&D spillovers are identified as the effect of R&D intensity in one region on R&D productivity in another. The results show that R&D spillovers are small and localized, with most benefits accruing to the region that spends R&D resources. The study also finds that R&D spillovers are more significant for regions that are closer to each other. The paper also considers the role of human capital in innovation and finds that higher levels of human capital are associated with higher levels of innovation. The study also examines the impact of distance on R&D spillovers and finds that the effect of R&D on innovation decreases with distance. The results suggest that R&D spillovers are strongest for regions within 300 km of each other. The study also considers the impact of historical population on R&D spillovers and finds that historical population is a good instrument for estimating R&D spillovers. The results suggest that R&D spillovers are more significant for regions with higher historical population. The study also finds that R&D spillovers are more significant for regions that are closer to each other. The results suggest that R&D spillovers are strongest for regions within 300 km of each other.