Governing by contract as a way to reduce crime? An impact evaluation of the large-scale policy of security pacts

In the last decades, governing by contracts, and in particular security contracts and pacts, has been increasingly promoted as a principal means of advancing crime prevention and governing security issues. Security pacts are a form of contract in which various institutional actors declare publicly to approve a common line of action, and/or taking mutual commitments and measures to prevent and control crime and disorder. Unfortunately, impact evaluations of the policy outcomes of security pacts are lacking. In this article, we aim to provide the first rigorous evaluation of the impact of a large-scale policy based on security pacts (involving around 12 million people) on various types of crimes in Italy. We built an ad hoc macro-level panel dataset of the 103 Italian provinces, with indicators covering a period spanning between 2004 and 2013. We applied generalized difference-in-difference models exploiting variation in the time and place in which the policy was adopted. Results indicate that security pacts had a limited impact on crime one year after the adoption, but significantly reduced thefts and micro-criminality two years after the adoption. We also found evidence of heterogeneous effects along province population size, with the largest effects in the larger provinces and null effects in the smaller ones. These findings are robust to a number of different sensitivity checks.