CRIME AND ECONOMIC PERFORMANCE. A CLUSTER ANALYSIS OF PANEL DATA ON ITALY'S NUTS 3 REGIONS

Crime is a complex phenomenon which needs to be investigated at appropriate disaggregate temporal and territorial levels of analysis. The specic issue addressed in this paper concerns the possibility of classifying Italian provinces in groups by using a new methodology that combines cluster analysis with panel time series data regarding a wide range of crime indicators, economic performance indicators and other socio-demographic variables. Our main contribution is to show that crime is not inextricably linked to geographic location as is usually believed. From this point of view, the position of the Italian Mezzogiorno has two facets; one is found in those relatively not affluent provinces which are resistant to random street and organized crime while the other facet is encountered in those provinces which are caught in a vicious circle where increasing criminal activity and weak economic performance feed on each other to undermine the security of the population. The pattern for the Center-North of Italy is much more varied and composed of a series of four clusters. In particular, there is a relatively large cluster of provinces which should be considered at risk because of an increasing value added per capita, high rates of population replacement and female employment; all conditions that might attract and encourage criminal activity.