Geographical proximity and digitalization: an analysis of productivity spillovers in agri-food sector

Abstract Geographical proximity can play an important role in shaping productivity spillovers within the agri-food sector, influencing the interactions among firms. This study analyses how spatial proximity and connectivity affect the productivity gap of agri-food firms. The empirical analysis relies on micro-data from the Orbis database, covering about 1,960 agri-food companies located in Southern Italy (Apulia region). Firm performance is measured through labour productivity (value added per employee), and the dependent variable is defined as the productivity distance from the local top performer in the same sector and province. Four distance metrics are computed linear, road, travel time, and altimetric distance—to capture different dimensions of spatial proximity. The econometric specification is based on cross-sectional OLS regressions with robust standard errors, while the quality of territorial connectivity is included as an independent covariate. As a robustness check, quantile regressions ( τ = 0.25, 0.50, 0.75) on the linear distance are reported in the Appendix, confirming the consistency of the main results. Overall, the analysis confirms that geographical proximity remains the main channel for the transmission of productivity spillovers. On the contrary, the average connection speed (fixed/mobile) does not show an appreciable impact on productivity gaps, a result consistent with the productivity paradox, without skills, reorganization and intangible capital, digital investments do not automatically translate into performance. These results provide useful insights for agri-food policy makers, particularly regarding the role of mobile communication networks in rural areas.