In the environmental valuation literature, a number of criteria have been set out for selecting among suitable studies for benefits transfer. The criteria focus on the environmental goods involved, the sites in which the goods are found, their beneficiaries and the quality of the study. Although benefits transfer has been discussed and applied extensively, especially in the United States, very little published evidence exists about its validity. The very few studies that have been carried out show negative outcomes regarding the validity of benefits transfer; in the case of contingent valuation (CV), even under the seemingly ideal circumstances where two identical questionnaires were used to value an identical good at two different sites at the same time. These results indicate that more research is needed to see how much control is necessary for valid benefits transfer.

In various publications the importance of evaluating similar goods and services at different sites when aiming at transferring their values has been emphasised, but not the importance of evaluating similar values within CV studies, i.e. different degrees of use and non-use values. The transfer of CV benefit estimates is expected to become more complicated in those cases where substantial off-site non-use values are involved. In this paper, we present the results of the third study to date which statistically tests the validity of environmental benefits transfer based on CV surveys and the first study which tests whether accounting for market size and use and possible non-use values associated with distance-decay effects affects the validity of transferring both benefit values and functions. Testing for so-called distance-decay effects in CV studies helps us to define the relevant population that benefits from the environmental goods involved (market size) and also provides important information about the possible non-use value attached to these goods. The studies included in the benefits transfer exercise comply with all the other selection criteria outlined in the benefits transfer literature.

Contrary to expectations, accounting for these effects does not make a difference with respect to the validity of the benefits transfer exercise: we find that benefit functions are not transferable. Moreover, whereas previous research showed that the equality of benefit function estimates does not guarantee equality of benefit estimates, we found that the reverse does not necessarily hold either. So, the transferability of benefit functions does not hold, even though we accounted for what we consider important criteria for succesful benefits transfer, namely market size and use and non-use values associated with distance-decay effects. Given the small data set used in the analysis, the results from this first single study cannot be generalised. The results do suggest that the current criteria for valid benefits transfer are not sufficiently rigid in order to make the exercise a viable input in environmental decision-making.