Home IGIER  

 

The Dark Side of Networks

       

PRIN Grant
Head of Bocconi Research Unit
Fernando Vega Redondo


Abstract

We study six aspects for which social networks amplify and propagate inefficient failures through local interactions, because agents do not internalize the negative externalities that their local behavior have, and the network works as a multiplier in the propagation of failures, or as a constraint in the flow of information. Each project is theoretical, but has empirical applications.

First, in a context of spread of an infection, i.e. a real disease or the diffusion of a bad habit, the endogenous adaptation of agents, that adjust their local network in response to the risk of being infected, can have the perverse effect of increasing the infectiveness of the disease. We find empirical evidence from the Italian National Bovine database.

Second, we want to conduct a lab experiment to see what are the channels for the diffusion of dishonest and anti-social behaviors, controlling for transmission of behavior that is either vertical (i.e. intergenerational) and horizontal (i.e. from peers and authority figures). We will do this in a High School in Siena, using as subjects students, their parents and their teachers.

Third, in supply chains of productions, the negative behavior of a critical mass of nodes can have avalanche effects on the aggregate trend of the whole society. We study the interactions of firms as games with local negative externalities and complementarity effects, where both action and interaction are endogenous. The model yields interesting predictions as the contagious effects act as a “technological trap” blocking economic development.

Fourth, we study a general equilibrium model of a complex economy where the network interaction among the different firms determines the individual payoffs and the overall performance. We have access to a complete panel data on all the transactions among (50 millions) Spanish firms, from the Spanish Tax Agency. With the combination of this “big data” and the theoretical model, the objective is to design an effective tool for tax inspection and the fight against tax fraud.

Fifth, in a context of informal mutual insurance, the more successful members of a social network must help the least successful members that they are linked to. So, resources redistribution within the network can be characterized like an ‘informal’ tax, and when this ‘tax’ is too high it can result in negative incentives for the individuals on their effort provision. We look for empirical confirmation from lab experiments in Tanzania.

Sixth, in several aspects of human behavior, opinions tend to diverge over time, generating the so called processes of polarization: people form endogenously their social contacts in a way that segregate them, and so different segregated communities converge to different or even contrasting opinions and norms. We study the interplay between network adjustment and opinion change that reinforces each other, being detrimental to social cohesion and social stability.

 
 
 

 



Last updated April 24, 2019