Exploring Benefit Transfer: Disamenities of Waste Transfer Stations

Abstract

Abstract. The benefit transfer method was developed as an alternative way to value externalities using values from studies of similar circumstances, carried out at similar sites somewhere else, given the challenges and high costs inherent in assessing the actual cost. Specifically, in order to test the performance of the benefit transfer method, employing hedonic price models, this study focused on estimates of disamenities associated with waste transfer stations at four different cities in Israel. The sites were intentionally selected to represent a variety of circumstances. We transferred the estimated benefit function from a ''study site'' to a ''policy site''. The goodness of fit was examined by comparing the calculated value, with actual data from the policy site. To test the sensitivity of the benefit transfer function to socioeconomic and housing characteristics, it was repeatedly applied to different sets of observations. The findings suggest that a relatively large number of alternative benefit functions are transferable. Statistical inequality outcome regarding the degree of similarity between samples does not unequivocally rule out the appropriateness of transferring environmental values across studies. Transfer errors varied between )21% and +29%, and the absolute average error for all transfers was 15.4%. Errors were lower for transfer between relatively similar cities in terms of size and location and between sub-samples that were similar in socioeconomic characteristics and housing type. However, when a site with very dissimilar data was involved, the average absolute error rose to 19%.

What can we learn from benefit transfer errors? Evidence from 20 years of research on convergent validity

“…10 Specifically, we adapt the kernel estimator for models with unordered discrete regressors described in Ouyang, Li and Racine [28]. Our estimation procedure recognizes that the true form of ( ) in equation (4) is unknown.…”

Disentangling the Pure Time Effect From Site and Preference Heterogeneity Effects in Benefit Transfer: An Empirical Investigation of Transferability

“…Brouwer and Bateman 2005;Bateman et al 2011), while others compare transferred predicted values with the predicted value at the policy site (e.g. Barton 2002;Rozan 2004;Eshet et al 2007). We follow the latter approach, since it reasonably compares two corresponding values; otherwise we could have additional TE from comparing the simple mean and the predicted value of the same study site (e.g.…”

“…Furthermore, the required level of precision can vary depending on the context that the BT is used. Previously BT has been used in many different contexts, including health (Barton 2002;Rozan 2004;Brouwer and Bateman 2005;Smith et al 2006;Brouwer 2006), ecosystem services (Brouwer and Spaninks 1999;Baskaran et al 2010), water quality (Muthke and Holm-Mueller 2004;Brouwer and Bateman 2005), private goods (Kealy et al 1988), recreational areas (Loomis 1992;Downing and Ozuna 1996;Kirchhoff et al 1997;Teisl et al 1995) and amenities associated with house prices (Chattopadhyay 2003;Eshet et al 2007). …”

Scaling Up Ecosystem Services Values: Methodology, Applicability and a Case Study

“…In response to this need there is now a sizeable literature that tests the accuracy of value transfer. Rosenberger and Stanley (2006) and Eshet et al (2007) provide useful overviews of this literature. Transfer errors are generally expressed as the Mean Absolute Percentage Error (MAPE), which is defined as the difference between observed value and predicted value divided by the observed value.…”

“…The study shows that the accuracy of value transfer relies heavily on the similarity of study sites. Eshet et al (2007) examine the accuracy of transferring values for the disamenity of housing locations close to waste transfer stations between four cities in Israel. Value functions derived from separate hedonic pricing studies are used to transfer values for each site.…”