When risky decisions generate externalities

Many risky decisions generate either positive externalities (e.g., opening a small business, engaging in research and development, vaccinating) or negative externalities (e.g., smoking, reckless driving, unprotected sex). I develop a new experimental framework for evaluating risk tolerance when risky decisions that generate externalities by modifying the allocation task of Gneezy and Potters (1997). In the ‘risk with externalities’ framework, for each unit increase in a risky allocation, an external cost or benefit is imposed upon a group member. I further vary the magnitude of the external effect to investigate the sensitivity of risk taking to the marginal external effect. Results indicate that risky decisions are driven by a combination of social and non-social factors. When individuals both impose and receive externalities, the preferred level of risk taking is lower if the externality is negative, regardless of the magnitude of the marginal external effect. Risk taking is then further affected by the magnitude of the marginal external effect. For individuals who only impose externalities, behavior is consistent with a disutility from imposing harm (but not a sensitivity to the magnitude of the external effect). For those who only receive externalities, individuals increase (decrease) their risky allocation when they are in a decision environment where they receive external benefits (costs). Additional treatments suggest that results are consistent with decreasing relative risk aversion plus an additional disutility from either imposing or receiving harmful externalities.