Although the positive role of transportation infrastructure in regional economics is well recognized, the impact could be spatially differentiated and examining such distributional effect of transportation infrastructure requires data at the disaggregated level. In this paper, I adopt the novel air pollution and thermal anomalies remote sensing data to measure local industrial activities by exploiting the opening of highspeed railway (HSR) in China. Using staggered difference-in-difference method, I show that the opening of HSR decreases local industrial activities by 6%, mainly due to the disappearance of small firms. In addition, I construct a new dataset using the remote sensing data to show that HSR decreases firm entry rate and increase firm exit rate for small firms and input-intensive industries are more affected by HSR. As suggested by a stylized general equilibrium model, HSR lowers search cost between industrial buyers and suppliers and induces only the more productive firms to stay in the market.
Measuring Real-time Economic Activities using Air Pollution Data: Evidence from COVID-19 in China, with Shanjun Li, Deyu Rao
Understanding real-time impacts of economic shocks at a granular level is crucial in designing effective policies to mitigate undesirable consequences, especially in emergencies like COVID-19. However, government statistics on GDP usually exhibits significant delays and lack the granularity needed to measure heterogeneous impacts. In addition, the credibility of official statistics in some developing countries also worth discussion. In this paper, we develop a pollution decomposition model to create a real-time proxy of GDP. As suggested by our GDP proxy, although Chinese economy experienced a V-shape recovery after the pandemic, the recession is more severe in many cities, compared with government GDP statistics.
Does Family Wealth Change Affect Education Choices? Evidence From the Housing Market in China
This paper measures the wealth effect of rising housing value on high school and college enrollment in China using the China Family Panel Studies data from 2012 to 2016. I find evidence that rising housing wealth increases the probability of enrollment in both high school and college education for low-income families: a 100,000 yuan increase in housing wealth increases high school enrollment by 4.65 percent and college enrollment by 7.34 percent. By seperating the pure wealth effect and borrowing constraint, I argue that the wealth eect in rural areas mainly comes from the relaxation of borrowing constraint due to the existence of larger network effect ”Guanxi” in rural China. These estimates imply that more policies targeting low-income families are required to provide more higher education opportunities in developing countries including China.