Accounting for unobserved individual heterogeneity in spatial stochastic frontier models: the case of Italian innovative start-ups

Spatial correlation and individual effects are two key sources of heterogeneity that should be handled in empirical applications investigating the productive performance of firms, especially when dealing with start-up activity. Therefore, in this work, we propose a fixed-effects spatial autoregressive stochastic frontier model for unbalanced panel data, we test the finite sample properties of our spatial estimator and we provide an empirical application on Italian innovative start-ups. The results of our analysis indicate that Italian start-ups are characterised by a significant level of global spatial dependence, especially overall and among firms belonging to the information and communication sector.