Risky Behavior Initiation with Behavioral agents and Social Interactions

Social interactions are widely recognized as critical in initiating many habit-forming processes, especially regarding substance use. Nevertheless, little attention has been given to the behavioral biases of individuals as a potential driving factor in such consumption. This paper presents a theoretical framework to study the incidence of these behavioral biases in health risky behavior initiation. In particular, we show that friends' biases in self-control failure are an essential part of the endogenous peer effects spillovers. We illustrate these results using Add Health data which contains unique information on friendship networks among U.S. teenagers. Empirically, we find that the contribution of friends' self-control in student's risky behavior is not negligible: a one standard deviation increase in friends' self-control leads to a reduction of 0.05 standard deviations in risky health behavior. Additionally, we can identify the key player in this setting. This student does not always coincide with the student with the highest (not micro-founded) centrality measures or most active substance user. We find that targeting the key player leads to a higher reduction in optimal aggregate health risky behavior than a policy targeting the most active player