STEFANO IANDOLO | Projects

SPATIAL PROXIMITY AND FIRM PERFORMANCES. EMPIRICAL EVIDENCE FROM ITALY, TURKEY AND TUNISIA.

The first issue we want to analyse is the relationship between agglomeration, innovation and firms’ productivity. Analysis of micro and macro factors which affect the efficiency conditions of firms are crucial in this context and they entail the consideration of proximity between firms, agglomeration indicators and R&D spillovers. The questions are: do firms localised in clusters of production exhibit a higher output? How far concentration of intangible assets, patents and R&D of co-located firms in the same cluster increase it? We also explore the complementarity between domestic and foreign firms. Firms should benefit from the experience of other firms in the vicinity, especially from the one of large foreign multinational firms. So we ask how localisation of firms nearby multinationals operating in the same localised cluster would contribute to develop their productivity and would allow innovation to circulate.Hence, the analysis for each economy aims to provide a measure of spillovers on output from geographical and sectorial clustering of firms and from their innovation. In addition to considering innovation measures at firm level, we build specific indexes of innovation activity at territorial level (regions for Turkey, governorates for Tunisia, provinces for Italy). We also use indicators of innovation performed by domestic and by foreign multinationals at the spatial level of analysis adopted. To sum up, we try to capture regional and sectoral spillovers from agglomeration of activities, from foreign firms and from innovation performers (both domestic and foreign) in the sector and in the spatial unit under analysis. In all the three studies we try to include explicit regional variables and to capture regional spillovers due to agglomeration economies and to innovation spillovers. In this way, we directly contribute to the wide literature on productivity spillovers from agglomeration economies, as well as to the literature on localized knowledge spillovers from innovation. Hence, we check for spillovers between firms taking place within regions, controlling for certain regions being more conducive to generating productivity growth (base levels of regional attractiveness, e.g. value added per head). We adopt different spatial unit of analysis and different proxies for innovation across the three country studies. However, we use the same estimation strategies: panel estimates for output controlling for time fixed effects. Simultaneity and endogeneity is addressed using system Generalized Method of Moments (GMM) dynamic panel estimation techniques. This methodology allows us to distinguish the direction of the nexus clustering and productivity and to focus on whether more regional clustering lead to higher productivity, ruling out the other direction of causality, i.e. that higher productivity leads to more regional clustering. It also allows to understand whether there are spillovers between firms taking place within regions, controlling for regional attractiveness, to avoid to catch a different relationship i.e. that certain regions more developed are more conducive to generating productivity growth. The methodology we use, GMM system, takes care of endogeneity of input by creating instrumental variables from existing variables. We also compare as robustness check other estimations methods (OLS, Fixed effects, Levinshon and Petrin, Olley and Pakes).