Working Paper: CEPR ID: DP13120
Authors: Carlos Daz; Eleonora Patacchini; Thierry Verdier; Yves Zenou
Abstract: This paper presents a new theory of crime where leaders transmit a crime technology and act as a role model for other criminals. We show that, in equilibrium, an individual's crime effort and crime decisions depend on the geodesic distance to the leader in his or her network of social contacts. By using data on friendship networks among U.S. high-school students, we structurally estimate the model and find evidence supporting its predictions. In particular, by using a definition of a criminal leader that is exogenous to the network formation of friendship links, we find that the longer is the distance to the leader, the lower is the criminal activity of the delinquents and the less likely they are to become criminals. This result highlights the importance of the closeness centrality of the leaders in explaining criminal behaviors. We finally perform a counterfactual experiment that reveals that a policy that removes all criminal leaders from a school can, on average, reduce criminalactivity by about 20% and the individual probability of becoming a criminal by 10%.
Keywords: crime; leaders; social distance; criminal decision; closeness centrality
JEL Codes: C31; D85; K42
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Distance to criminal leaders (K14) | Criminal activity of delinquents (K42) |
| Distance to criminal leaders (K14) | Likelihood of becoming a criminal (K14) |
| Distance to criminal leaders (K14) | Influence of leaders (M54) |
| Distance to criminal leaders (K14) | Criminal activity (weaker influence at distance) (K42) |
| Removal of criminal leaders (P37) | Average criminal activity (K42) |