From creativity to innovativeness : micro evidence from Italy

In this paper I assess the existence, and the magnitude, of technological externalities in the form of creativity spillovers that affect individual firms' innovative intensity. Relying on a large sample of Italian manufacturing firms, I first estimate a knowledge production function through a zero-inflated beta regression model and a generalized Tobit model. From these, I extract the residuals, which represent the unexplained part of the actual observed share of innovative sales, namely innovativeness. Then, I regress such a measure of firm innovativeness on a set of occupation-based, as well as density-based, indicators of creativity at the NUTS3 level, while controlling for firm localization, size and industry. I also control for endogeneity and non-linearity by estimating a two-stage least squares model and a generalized additive model respectively. My estimates show that: (i) there is a positive and highly statistically significant effect of creativity on innovativeness; (ii) the effect of creativity on actual innovative sales is weak, while I find a stronger effect played by the availability of R&D labour within the firm; (iii) occupation-based measures of creativity outperform education-based measures of human capital; (iv) when controlling for the education content of jobs, firms' innovativeness is affected more by the local availability of non-graduated creative workers than of graduated ones; (v) rather, relying on NUTS3 regional data, I find that a higher local availability of graduated creative workers affects the invention intensity of a city; (vi) the relationship between firm innovativeness and the local density of creative people is U-shaped, so that proximity-based knowledge externalities emerge only after a certain density threshold is reached, this occurring typically in larger urban areas, typically hosting design and service-based creative industries. From the policy point of view, increasing the availability of creative jobs and people can help regions and cities to be more innovative, especially in the absence of large R&D departments and formal agreements with external partners. In this respect, my results are in line with the literature on innovative milieux, where social learning phenomena, rather than formal R&D activities, help explaining the processes of knowledge creation and diffusion within and between firms, clusters and territories.