Travel to breathe the fresh air? Big data evidence on the short-term migration effect of air pollution from China

This study uses unique daily city-pair population flow intensities from Tencent migration big data to estimate the effect of air pollution on population outflow. Our identification strategy relies on high-dimensional fixed effects models to isolate city and city-pair level potential confounders varying or invariant over the year. We document that people respond to air pollution by traveling to cities with better air quality. The elasticity of population outflow intensity to the Air Quality Index (AQI) is estimated as around 0.009, suggesting that increasing AQI by 100 points raises the population outflow intensity by about 1.5%. The instrumental variables method leads to a doubled elasticity of 0.018. We also explore several heterogeneity effects, showing that the short-term emigration effect increases as the air quality deteriorates, can occur within two days before and after, is larger on national holidays, and is smaller for cities with unique tourism resources. We finally reveal that people flee from polluted cities by car and train rather than by plane, implying a short- or medium-distance traveling effect of air pollution. This study thus provides clear evidence that air pollution drives away population and suggests that air pollution control can produce social gains in addition to those related to health and properties.