The mafia is often considered one of the main causes of the “questione meridionale”, the historical social and economic gap between Northern and Southern Italy. The most visible aspect of the mafia in the local economic environment is the “pizzo”, a slang term for protection money paid to the mafia under extortion. Recent estimates show that more than two-thirds of Sicilian firms pay the pizzo and the average monthly payment is approximately 880 euros (this is higher than the average tax burden).
Why aren’t these firms able to coordinate and jointly refuse to pay the pizzo? Previous literature has conjectured that the payment of the pizzo can be rationalized in a context in which lack of trust makes it difficult for businesses to share information and coordinate against the mafia. The aim of this project was to test this conjecture and provide related policy indications by means of a field experiment. Specifically, we wanted to test and evaluate the effectiveness of a policy aimed at increasing the level of generalized trust and, through this channel, at inducing entrepreneurs to be more vocal in opposing the payment of pizzo.
To this purpose, we exploit a unique phenomenon of rebellion that has emerged in Italy: the creation of Addiopizzo, an NGO that unites firms that voluntarily and collectively refuse to pay protection money. A crucial element for success of this action is that a firm’s anti-pizzo stance is made public, as a signal to other firms of lower costs of rebelling, thus triggering a virtuous contagion effect. Nonetheless, we have collected preliminary evidence that:
1. A large number of firms in the Addiopizzo network do not publicize their membership
2. Firms in the network underestimate the number of Addiopizzo members in their neighborhood
The field experiment was supposed to test whether increased information and easier coordination among Addiopizzo firms have positive effects on engagement in the association and, eventually, on the perception of the pizzo among customers, residents and other firms.