This paper investigates the degree of comovements in quarterly Italian time series of sectoral output. A recently developed multivariate technique for the empirical analysis of long-run, cyclical and seasonal comovements is used in the context of a multisectoral real-business-cycle model augmented with persistent seasonal shocks in productivity. Our empirical results emphasize the role of input–output relations in the propagation mechanism and indicate that sectoral outputs have a relatively low number of common stochastic trends, in conflict with the hypothesis of independent productivity shocks. In contrast, stochastic seasonals seem to move idiosyncratically. Furthermore, our findings suggest that the theoretical model should be extended to allow for deterministic seasonal shifts in preferences.
