Trans-boundary air pollution spillovers: Physical transport and economic costs by distance

The economic costs of trans-boundary pollution spillovers versus local effects is necessary to evaluate centralized versus decentralized environmental policies. Directly estimating these for air pollution is difficult because spillovers are high-frequency and vary with distance while economic outcomes are usually measured with low-frequency and local pollution is endogenous. We develop an approach to quantify local versus spillover effects as a flexible function of distance utilizing commonly-available pollution and weather data. To correct for the endogeneity of pollution, it uses a mixed two-stage least squares method that accommodates high-frequency (daily) pollution data and low-frequency (annual) outcome data and can improve efficiency. We estimate spillovers of particulate matter smaller than 10 mu g (PM10) on manufacturing labor productivity in China. A one mu g/m(3) annual increase in PM10 locally reduces the average firm's annual output by CNY 45,809 (0.30%) while the same increase in a city 50 km away decreases it by CNY 16,248 (0.11%). This effect declines rapidly to CNY 2847 (0.02%) for an increase in a city 600 km away and then slowly to zero at 1000 km. The results suggest the need for supra-provincial environmental policies or Coasian prices quantified under the approach.