Hedonic property models are commonly applied in the environmental economics literature to estimate values of environmental amenities or hazards. Most hedonic property models are estimated using linear regression techniques where the coefficient on the environmental variable of interest is the "marginal implicit price." Linear regression estimates one coefficient for the entire distribution of the dependent variable, and thus in hedonic property models a single marginal implicit price. In contrast, quantile regression estimates a range of marginal impacts for different quantiles of the distribution for the dependent variable, consequently providing a significantly more "complete picture" of the true impact of the explanatory variable (Koeneker and Hallock, 2001). We contribute to the existing hedonic property literature by estimating the impact of repeated wildfires on house prices in Southern California using quantile regression. We find that the impact of a wildfire differs significantly across the distribution of house prices, with estimated coefficients varying as much as 73% from the 25th quantile relative to the 75th quantile. We also find that OLS results under-estimate impacts relative to the median quantile, yet over-estimate impacts for lower quantiles. Our results indicate that a quantile regression approach can provide policymakers and researchers more information about the marginal implicit price in hedonic models as it relates to the distribution of the dependent variable.
- Hedonic property models
- Quantile regression
ASJC Scopus subject areas
- Geography, Planning and Development
- Nature and Landscape Conservation
- Management, Monitoring, Policy and Law