The Non-Linear Effect of Daily Weather on Economic Performance : Evidence from China

The paper examines the non-linear relationship between weather and economic performance. Using within-county variation in weather between 1996 and 2012, we find both daily temperature and precipitation have asymmetric effects on economic performance. A day with temperatures over 20◦C significantly lowers the county-level GDP, while the effects of low temperatures are not significant. Reduction in low-intensity precipitation day (within (0,20]mm) and increase in zero precipitation day also significantly decrease the county-level GDP, but the high-intensity precipitation day has no significant effect. Additionally, agricultural sector is more vulnerable to weather extremes, while the effects of weather extremes on non-agricultural sector are not significant. Outputs of grains, cotton and oil crops are significantly reduced by extremely high temperatures and extreme precipitation. The agricultural sector is the main channel for weather extremes act on overall economic performance. We also find evidences that economic productions show adaptation to weather extremes: counties have experienced more weather extremes present better adaptation; crop production in agricultural counties is less affected by weather extremes; penetration of air conditioning mitigates the detrimental effect of overheat in rural areas