Understanding contextual heterogeneity in the outcomes of large-scale security policies: evidence from Italy (2007-12)

We develop a theoretically informed quantitative evaluation of the large-scale policy of security pacts (SPs) in Italy (2007-2012). In particular, we show how Matland's theoretical model of policy implementation (1995) may help clarify the contextual heterogeneity present in the outcomes of security policies. We build and use a macro-panel dataset in which we exploit variations in the temporal and geographic adoption of SPs to develop a dynamic analysis of the contextual factors that influence the probability of adopting a large-scale security policy at the sub-national government level. We apply a rigorous counterfactual design to assess the effectiveness of SPs in reducing theft and robbery rates. We find that the probability of adopting SPs is positively affected by coalitional strength but not political orientation. In contrast, the effectiveness of SPs in reducing crime relates more to the political orientation of the ruling parties than to coalitional strength.