This paper reports results from two stated preference (SP) experiments valuing aircraft noise around Manchester and Lyon airports. It builds substantially upon the authors' earlier papers (Bristow et al., 2004 and Wardmanet al., 2003). This paper incorporates modelled aircraft noise data for respondents' addresses which has enabled us to derive a value of noise (Leq) and offer the following additional key developments. Firstly, a re-estimation of the SP models based on aircraft movements to include the current (or base) level of aircraft noise (Leq) which improved model fit in all cases. The models were then further modified to replace movements with Leq which again (in most cases) gave a slightly better fit than the original aircraft movements model. This model too was adapted to allow for the base level of noise through a power term. Whether based on movements or Leq the models were improved by allowing for the current level of noise and in seven out of eight models the unit value increases as the base level of aircraft noise increases. This evidence supports the case for a level effect. Secondly, threshold effects were examined in the aircraft movements model with thresholds at 45, 50 and 55 dB(A) tested. Contrary to our expectations, there was no support for the presence of threshold effects. It is possible that there are confounding effects at work here such that those who have higher (lower) values reside farther away (closer to) the airport. Thirdly, different noise indices were examined: Leq, Lmax (65 and 70) and SEL (65 and 70). The results provide some evidence that Leq provides a better account of how respondents value aircraft noise than rival indices. However, caution is urged here as respondents would not actually have experiencedthe different noise levels presented except insofar as they mapped onto variations by time of day or other time period. movements models were used to test a wide range of socio-economic, situational and perceptual variables to identify any impact on the values of noise. The second SP experimentwas more successful in identifying influential variables, some of which including income, probability of being at home, annoyance from aircraft noise, household size and a zero response to the contingent valuation question are common between Lyon and Manchester which is encouraging. Finally, weexamined the internal consistency of the results from the quality of lifeSP by examining the rating and values of different variables within it. Acomparison with results from other studies is also undertaken; although this is against the very small number of SP studies and selected hedonic pricing studies our results in terms of the unit value of change in Leq do not appear to be out of line with those identified elsewhere. For the covering abstract see ITRD E135582.
Abstract