Research

Abstract:

We provide a game-theoretical epidemiological model for the COVID-19 pandemic that takes into account that: 1) asymptomatic individuals can be contagious, 2) contagion is behavior-dependent, 3) behavior is determined by a game that depends on beliefs and social interactions, 4) there can be systematic biases in the perceptions and beliefs about the pandemic. We incorporate lockdown decisions by the government into the model. The citizens' and government's beliefs can exhibit several biases that we discuss from the point of view of behavioral economics. We provide simulations to understand the effect of lockdown decisions and the possibility of "nudging'' citizens in the right direction by improving the accuracy of their beliefs. In an extension, we allow citizens to protest these measures if they believe they are imposed for autocratic reasons.

Abstract:

Social interactions are widely recognized as critical to the initiation of many habit-forming processes, particularly concerning substance use. However, little attention has been paid to the behavioral biases of individuals' friends as a potential driving factor in such use. In this paper, we present a theoretical framework for investigating the role of these behavioral biases in the initiation of health-risk behaviors. In particular, we show that friends' biases in the form of self-control failures are an essential part of endogenous peer effects spillovers. We illustrate these results using data from Add Health, which contains unique information on the friendship networks of U.S. adolescents. Empirically, we find that the contribution of friends' self-control to students' health risky behavior is not negligible. In addition, we can identify the key player in this setting. This student is not always the student with the highest (non-microfounded) centrality measures or the most active player. We find that targeting the key player leads to a greater reduction in optimal aggregate health risky behavior than a policy targeting the most active player or a random player.

Abstract:

The increasing incidence rate of diagnosed ADHD has sparked debates about over-medicalization and misdiagnosis. We use data on the genetic predisposition for ADHD from the Add Health survey on U.S. schools to uncover relative standards in ADHD diagnosis. We estimate that students' ordinal rank in the genetic predisposition to ADHD among their same-gendered grademates has a positive, statistically significant, and substantial causal effect on ADHD diagnosis, holding students' own genetic predisposition to ADHD constant. This effect is mainly driven by boys, contributing to explain the observed higher rate of diagnosis of boys relative to girls for a given genetic ADHD predisposition.