CEP Discussion Paper No 1126 February 2012 Happiness as a Driver of Risk - Avoiding Behavior

Most governments try to discourage their citizens from taking extreme risks with their health and lives. Yet, for reasons not understood, many people continue to do so. We suggest a new approach to this longstanding question. First, we show that expected-utility theory predicts that ‘happier’ people will be less attracted to risky behaviors. Second, using BRFSS data on seatbelt use in a sample of 300,000 Americans, we document evidence strongly consistent with that prediction. Our result is demonstrated with various methodological approaches, including Bayesian model-selection and instrumental-variable estimation (based on unhappiness caused by widowhood). Third, using data on road accidents from the Add Health data set, we find strongly corroborative longitudinal evidence. These results suggest that government policy may need to address the underlying happiness of individuals rather than focus on behavioural symptoms.