Marginal values of traffic noise externalities from stated preference methods.

Although the valuation of the external effects of transport noise has been the subject of a large number of hedonic pricing studies, relatively few studies have applied stated preference (SP) methods to this problem. This paper reports on a novel application of SP to the valuation of road traffic noise by residents in Lisbon, Portugal. It is set within the context of households' choices amongst different apartments in the same block of flats with different levels of exposure to the main sources of traffic. Traffic noise is evaluated alongside a number of other housing attributes in the SP exercise. The advantage of the approach adopted, crucial in the context of attributes such as noise which are difficult to represent in SP exercises, is that the respondents were familiar with different levels of noise in different flats according to elevation and location relative to the main road. A CAPI survey was conducted amongst 412 households, generating 4944 observations for the modelling stage. In addition, internal and external noise measurements were taken at a large number of the sampled flats which allows relationships to be developed between values, perceptions and physical measures. The results reported in this paper are based on standard logit and mixed logit analysis of the SP responses. A large range of variables was examined to establish the degree of preference heterogeneity across the sample. These include current exposure to traffic noise, household income, averting behaviour, attitudes to traffic noise and household characteristics. Issues of functional form, such as improvements and deteriorations in traffic noise and dependency of the marginal value on the absolute level of noise, were also examined and found to be important. The noise variable contained in the models was expressed both using respondents' perceptions as measured on a rating scale and the equivalent physical measures taken in situ (Leq dB(A)). A comparison of the performance of these models was made. The paper concludes with a comparison of the SP based values against those obtained within the same study from the more conventional contingent valuation method and with values obtained from RP models based on households' actual apartment choices. For the covering abstract see ITRD E124693.