Mafia, Politics and Machine Predictions

• In Event: Causal Inference and Algorithmic Discovery for Criminology and Public Policy

Abstract

We present a machine learning approach to predict local governments infiltrated by mafias in Italy, based on the dismissal of city councils infiltrated by organized crime from 2001 to 2020. We successfully predict up to 96% of out-of-sample municipalities previously identified as infiltrated by mafias: Concentrating every year on a small subset of high-risk municipalities (i.e. 4.5%), our index correctly identifies infiltrated municipalities without being too costly on the side of false positives. Second, we flag high-risk municipalities identified as infiltrated by our index but never investigated by the Italian state, as a group of municipalities suitable for targeted investigations. We then apply this measure to investigate a research question requiring time-varying granular data on mafia influence on politics: whether redistributive policies encouraging economic growth in depressed areas (European Union transfers) affect mafia infiltration in politics. Employing a Difference-in-Discontinuities design, we find a substantial and lasting increase in the predicted risk of mafia infiltration (up to 14 p.p.), emphasizing the risks of delivering aid where criminal organizations can appropriate public funds. This index can be used by policymakers to improve the detection of organized crime infiltration in politics and by researchers interested in investigating ties between mafia and politics.

Authors

• Gian Maria Campedelli, Fondazione Bruno Kessler

• Gianmarco Daniele, University of Milan / Bocconi University

• Marco Le Moglie, Catholic University, Milan