Abstract:
We utilize region-day variation in actual vs. forecasted weather conditions (i.e. forecasting errors) to comprehensively investigate the effects of the first heat-health warning system in Japan. We find that heat alerts led to a 17% increase in heat stroke counts. An analysis of mechanisms suggests this is primarily driven by increased reporting, as opposed to potential "adverse" behavioral responses or substitution in health diagnoses. Four times as many heat strokes are detected in low-income neighborhoods compared to high-income neighborhoods. We further document evidence of avoidance behaviors (e.g. electricity usage), but find no evidence of improved (or worsened) mortality outcomes.