Revealed and stated preference valuation and transfer: a within-sample comparison of water quality improvements

Benefit transfer (BT) methods are becoming increasingly important for environmental policy,
but the empirical findings regarding transfer validity are mixed. A novel valuation survey was designed to obtain both stated preference (SP) and revealed preference (RP) data concerning river water quality values from a large sample of households. Both dichotomous choice and payment card contingent valuation (CV) and travel cost (TC) data were collected. Resulting valuations were directly compared and used for BT analyses using both unit value and function transfer approaches. WTP estimates are found to pass the convergence validity test. BT results show that the CV data produce lower transfer errors, below 20% for both unit value and function transfer, than TC data especially when using function transfer. Further, comparison of WTP estimates suggests that in all cases, differences between methods are larger than differences between
study areas. Results show that when multiple studies are available, using welfare estimates from the same area but based on a different method consistently results in larger errors than transfers across space keeping the method constant.