Abstract
Several scholars recently investigated international drug trafficking using social network analysis (Chandra and Joba 2015; Boivin 2014). These studies do not quantify the volume of drug flows among countries thus analyzing networks with binary links. This study estimates the size of the heroin flows from and to European countries. Then, it exploits recent advances in statistical models for social networks to identify the factors shaping heroin trafficking. It first assesses the influence of such factors on the presence of a link between any two countries. It then introduces the edge weights (i.e. heroin exchanged) in the analysis, moving forward in the analysis of drug trafficking networks. The methodology adopted allows for a better understanding of the generative processes and potential evolution of trafficking routes. The study shows that opportunity theory and social proximity are keys in explaining the direction and the size of heroin flows in Europe. Moreover, it demonstrates the benefits of including weights in the analysis of drug networks.