Environmental Engineering Reference
valuation. CM emphasises thus not only monetary valuation. Because CM is designed
to rank multi-attribute alternatives, it seems more suitable to value agricultural
landscapes and the countryside as a whole, which provide a multitude of joint
goods and services, than the typical one-dimensional CVM (Farber and Griner 2000) .
It allows obtaining values for specific characteristics of (landscape) scenarios rather
than one specific scenario. On the other hand, CM can be more complicated to
implement, and the selection and representation of the several attributes adds to the
design problem already associated with hypothetical surveys.
While the debate over the existing techniques for monetary valuation continues, the
search for new approaches remains. The method of benefit transfer is fairly new
and still a developing subject (see, for example, the special issue of Ecological
Economics on the state-of-the-art and science of benefit transfer, edited by Wilson
and Hoehn 2006) . Benefit or value transfer involves 'borrowing' monetary environ-
mental values estimated at one site (the study site) and applying them to another
(the policy site). The attraction of benefit transfer is that it avoids the cost of
conducting 'primary' studies whereby the benefits of natural or environmental
assets or the damages associated with degradation are measured with one or more
of the techniques described above (OECD 2002) .
Navrud (2001 ; see also Rosenberger and Stanley 2006 , and Spash and Vatn
2006) distinguishes two main approaches to benefit transfer: namely, unit value
transfer, and function transfer. Unit value transfer is the easiest approach and refers
to simply transferring the results of an environmental valuation study to the policy
site. The valuation estimates may be left unadjusted, or may be adjusted in some
way. Although transferring unadjusted estimates is undesirable, since the values
experienced in the study site may not be the same as in the policy site, it is widely
practised (OECD 2002) . A more sophisticated approach is to transfer the entire
benefit function. Thus, instead of transferring only the benefit estimates, a
researcher transfers a benefit function, substituting data on independent variables at
the new site. Both unit values and value functions can be based on multiple studies.
Meta-analysis can be used to take the results from a number of studies, synthesise
research findings, and analyse them in such a way that the variations in unit values
or value functions found in those studies can be explained (van den Bergh et al.
1997 ; Navrud 2001 ; Dalhuisen 2002) .
The advantage of benefit transfer is that it is both pragmatic and cost-effective
and therefore, the method is a very attractive alternative to expensive and time
consuming original research to quickly inform decision making (Garrod and
Willis 1999a) . To date, however, benefit transfer is a controversial issue in various
policies contexts. The major reason for this is that the accuracy of transferred
estimates is debated because validity tests show that the uncertainty in transfers
could be quite large, both spatially and intertemporally. This is especially the case
for complex goods, such as biodiversity in agriculture and landscape values.