The role of optimizing behavior in willingness-to-pay estimates for air quality

Revealed preference models used in nonmarket valuation have, at times, overlooked potentially important implications of the assumption of optimizing behavior in measuring individuals' values for amenities. Epple and Sieg (1999a) have developed a new empirical framework that can be adapted to illustrate this general point by using the spatial equilibrium conditions to identify and estimate preference functions that include local and environmental public goods. After outlining our modifications to their framework in the next section, we illustrate how it can be used to estimate the marginal willingness to pay for reducing air pollution in southern California. Estimates based on a hedonic model are also presented for comparative purposes. Our preliminary estimates of marginal willingness to pay (WTP) measures for reducing ozone (in 1990 dollars) from the spatial sorting model are between 9% and 32% smaller (depending on the county and air pollution measure used) than the hedonic estimates. These estimates move consistently with the averages computed from the hedonic model's point estimates for each housing sale in the thirty-six school districts included in this preliminary analysis. These comparisons are certainly not tests for differences across models. However, they do suggest recognizing the discreteness in local public goods, and site-specific amenities may well introduce an important new restriction, implied by the sorting equilibrium framework, to the measurement of consumers' values for improvements in air quality.