A Spatial Model of Air Pollution: The Impact of the Concentration-Response Function

We develop a spatial model to examine policies aimed at reducing ambient concentrations of fine particulates (PM2.5), with emissions from many sources that affect many population centers. Two alternative specifications of the relationship between PM2.5 concentration and health impacts from Krewski et al. are analyzed: log-linear, which implies downward-sloping marginal benefits of abatement; and log-log, which implies upward-sloping marginal benefits of abatement. A standard assumption would be that the greatest benefit from cleanup would occur in the dirtiest locations. We show, however, that for the log-log (but not log-linear) relationship, the largest risk reductions are achieved from abatement of pollution in the cleanest locations. Our model demonstrates that with a log-log relationship society should prefer lower emissions and lower pollution concentrations than if the relationship is log-linear. Our model also shows that an efficient abatement policy may substantially outperform a uniform pollution standard such as the National Ambient Air Quality Standards (NAAQS).