Abstract
The value of noise is often estimated as a constant percentage of house price depreciation per decibel of noise. We examine this assumption and determine to what extent self-selection influences noise values from spatial Hedonic Pricing (HP) and Stated Choice (SC) approaches, exploring key aspects of noise valuation, such as non-linearities, thresholds and stigma. The HP models show non-linear house price depreciation per decibel of aircraft noise, increasing from 0.40% per decibel at 45 decibel to 2.38% at 75 decibel. HP models applied after the total removal of aviation noise show one specific area, with significantly higher long- Term exposure to aircraft noise, to be subject to noise "stigma" or unalleviated noise depreciation. SC noise values in the same "noisy" area are considerably lower than neighbouring areas, attributed to self-selection and the resulting higher noise tolerance of local residents. The HP results show that the housing market does not reflect this increased noise tolerance of the residents. However, the nonlinearity of the HP noise values reflects the self-selection effects for observations with high aviation noise, and self-selection can also remain as a localised stigma after the removal of the effect. These findings have implications for stated preference applications and the representation of noise costs in policy.