Long-run dynamics and cyclical patterns of R&D and innovation: evidence from Italy

This paper investigates the long-run relationship between research and development (R&D) expenditure and patent-based innovative output in Italy, focusing on the role of institutional funding sources and territorial heterogeneity. Using a dataset considering Italian macro-areas (1995–2022) within a knowledge production function framework, we apply a Panel Autoregressive Distributed Lag (ARDL) model, based on the Pooled Mean Group (PMG) estimator, to estimate cointegrating relationships and short-run dynamics. The principal results show that business R&D expenditure has a significant positive long-run association with patenting activity. Conversely, both government and higher education R&D expenditures exhibit significant negative long-run coefficients, pointing to possible institutional misalignment and crowding-out effects. The medium-term cyclical analysis based on the Baxter-King filter indicates that higher education R&D tends to display the highest positive co-movement with patenting activity across macro-areas (although this synchronization is statistically significant only in the North-West), consistent with a delayed, structural contribution rather than an immediate effect. The major conclusion is that the effectiveness of R&D policies depends critically on institutional coordination, the timing of impacts, and regional characteristics. Policies should be redesigned to mitigate crowding-out by targeting earlystage research, recognize heterogeneous temporal dynamics, and account for territorial specificities to strengthen long-term innovation performance.