Abstract
I use the life satisfaction approach to value air quality, combining individual-level panel and high-resolution SO2 data. To avoid simultaneity problems, I construct a novel instrument exploiting the natural experiment created by the mandated scrubber installation at power plants, with wind directions dividing counties into treatment and control groups. I find a negative effect of pollution on well-being that is larger for instrumental variable than conventional estimates, robust to controls for local unemployment, particulate pollution, reunification effects and rural/urban trends, and larger for environmentalists and predicted risk groups. To calculate total willingness-to-pay, the estimates are supplemented by hedonic housing regressions.