16.5 Techniques to Measure Nonmarket Economic Values
enough to include houses with similar features over
a range of different levels of the environmental
amenity in question. As with other indirect methods, a hedonic model can be developed using secondary data. For the model to be successful, the
data must include enough variations of combinations of characteristics. The price of a particular
model as a function of the quantities of its various
characteristics is a hedonic price function. The partial change in the hedonic price function with respect to a change in one characteristic is the implicit price of that characteristic. Thus a change in
the hedonic price function with respect to a change
in the level of water quality would be interpreted
as the additional cost necessary for a consumer to
purchase a unit of water quality.
Hedonic pricing results can be difficult to estimate and to interpret, because the comparisons
across purchase decisions assume that the differences in house prices can be attributed to different
levels of the amenity. Since the comparisons are
interpersonal, there are potentially any number of
reasons why consumers may choose to make the
purchases that they do. The preference functions of
different individuals may not be comparable, but
the analyst is not able to observe sufficient information about other factors that may influence an
individual's choice. Freeman (1993) reviews many
of the problems encountered with estimation of hedonic models.
16.5.3 Benefits Transfers
A third type of valuation method is the benefits
transfer. In reality, a benefits transfer is not a proper
valuation method, but rather a procedure that uses
valuation estimates from other study sites to apply
to a given policy site. One reason for performing a
benefits transfer is to make a first-pass estimate on
the potential range of values that might be attributed to a specific policy proposal in order to determine whether a full valuation study is warranted.
Few economists now suggest that estimates from
benefits transfer are reliable enough to inform a final decision for which the potential gain or loss of
nonmarket values might be very great or in situations in which the decisions are not reversible
(Downing and Ozuma, 1996; Kirchoff et al., 1997).
Rollins and Ivy (1997) provide a practical example of the steps involved in a benefits transfer approach to assessing the nonmarket values of a policy proposal to construct a dam and irrigation
system in a prairie and rangeland ecosystem in
North America.
The concept of benefits transfer evolved from an
239
initially naIve view that valuation measures of a
given ecosystem-based good or service are applicable to situations other than the one in which the
estimates were measured. This idea was implicitly
adopted by noneconomists who knew enough about
nonmarket values to realize the significance of
omitting them from an analysis, but not enough
about the underlying concept of economic value to
understand that, in general, estimates are not transferable from one policy scenario to another. At a
time when budgets for environmental impact assessments and cost-benefit analyses were weighted
toward engineering studies and the notion of nonmarket valuation was still considered experimental, a benefits transfer approach may have been considered one of the few means to attract attention to
the importance of nonmarket values. Over time,
benefits transfers came to be seen as a fast and inexpensive substitute for nonmarket valuation.
Academic economists have generally agreed that
benefits transfer is not valuation and is an inferior
means to estimate nonmarket values. However,
most are also aware that benefits transfers will continue to be used in applied policy, due to the speed
and cost relative to the cost and time involved for
proper valuation studies. Therefore, much research
focuses on developing criteria to determine the
types of circumstances for which benefits transfer
is more or less likely to be a satisfactory alternative to valuation (AERE, 1992; Atkinson et aI.,
1992; Boyle and Bergstrom, 1992; Brookshire and
Neill, 1992; Desvousges et al., 1992; Loomis,
1992; McConnel, 1992; Walsh et aI., 1992; Smith,
1993; Boyle et aI., 1994; Kask and Shogren, 1994;
Pearce et al., 1994; Hagler Bailly Consulting, Inc.,
1995; Loomis et al., 1995; VandenBerg etal., 1995;
Bergstrom, 1996; Rollins and Ivy, 1997; Desvousges et al., 1999). One common conclusion from
most of these studies is that there is no one set protocol for conducting benefits transfer. That is, there
can be no cookbook approach because the circumstances vary so widely frolJl one policy situation to
another.
16.5.4 Choosing the Technique to Fit
the Situation
There are often obvious circumstances that suggest
the most suitable technique for a given valuation
problem. For example, contingent valuation is the
only accepted method available for measuring existence values. We may rule out a benefits transfer
approach to provide information for a benefit-cost
analysis of a proposed policy decision if the possible cost of an error is very great relative to the po-
enough to include houses with similar features over
a range of different levels of the environmental
amenity in question. As with other indirect methods, a hedonic model can be developed using secondary data. For the model to be successful, the
data must include enough variations of combinations of characteristics. The price of a particular
model as a function of the quantities of its various
characteristics is a hedonic price function. The partial change in the hedonic price function with respect to a change in one characteristic is the implicit price of that characteristic. Thus a change in
the hedonic price function with respect to a change
in the level of water quality would be interpreted
as the additional cost necessary for a consumer to
purchase a unit of water quality.
Hedonic pricing results can be difficult to estimate and to interpret, because the comparisons
across purchase decisions assume that the differences in house prices can be attributed to different
levels of the amenity. Since the comparisons are
interpersonal, there are potentially any number of
reasons why consumers may choose to make the
purchases that they do. The preference functions of
different individuals may not be comparable, but
the analyst is not able to observe sufficient information about other factors that may influence an
individual's choice. Freeman (1993) reviews many
of the problems encountered with estimation of hedonic models.
16.5.3 Benefits Transfers
A third type of valuation method is the benefits
transfer. In reality, a benefits transfer is not a proper
valuation method, but rather a procedure that uses
valuation estimates from other study sites to apply
to a given policy site. One reason for performing a
benefits transfer is to make a first-pass estimate on
the potential range of values that might be attributed to a specific policy proposal in order to determine whether a full valuation study is warranted.
Few economists now suggest that estimates from
benefits transfer are reliable enough to inform a final decision for which the potential gain or loss of
nonmarket values might be very great or in situations in which the decisions are not reversible
(Downing and Ozuma, 1996; Kirchoff et al., 1997).
Rollins and Ivy (1997) provide a practical example of the steps involved in a benefits transfer approach to assessing the nonmarket values of a policy proposal to construct a dam and irrigation
system in a prairie and rangeland ecosystem in
North America.
The concept of benefits transfer evolved from an
239
initially naIve view that valuation measures of a
given ecosystem-based good or service are applicable to situations other than the one in which the
estimates were measured. This idea was implicitly
adopted by noneconomists who knew enough about
nonmarket values to realize the significance of
omitting them from an analysis, but not enough
about the underlying concept of economic value to
understand that, in general, estimates are not transferable from one policy scenario to another. At a
time when budgets for environmental impact assessments and cost-benefit analyses were weighted
toward engineering studies and the notion of nonmarket valuation was still considered experimental, a benefits transfer approach may have been considered one of the few means to attract attention to
the importance of nonmarket values. Over time,
benefits transfers came to be seen as a fast and inexpensive substitute for nonmarket valuation.
Academic economists have generally agreed that
benefits transfer is not valuation and is an inferior
means to estimate nonmarket values. However,
most are also aware that benefits transfers will continue to be used in applied policy, due to the speed
and cost relative to the cost and time involved for
proper valuation studies. Therefore, much research
focuses on developing criteria to determine the
types of circumstances for which benefits transfer
is more or less likely to be a satisfactory alternative to valuation (AERE, 1992; Atkinson et aI.,
1992; Boyle and Bergstrom, 1992; Brookshire and
Neill, 1992; Desvousges et al., 1992; Loomis,
1992; McConnel, 1992; Walsh et aI., 1992; Smith,
1993; Boyle et aI., 1994; Kask and Shogren, 1994;
Pearce et al., 1994; Hagler Bailly Consulting, Inc.,
1995; Loomis et al., 1995; VandenBerg etal., 1995;
Bergstrom, 1996; Rollins and Ivy, 1997; Desvousges et al., 1999). One common conclusion from
most of these studies is that there is no one set protocol for conducting benefits transfer. That is, there
can be no cookbook approach because the circumstances vary so widely frolJl one policy situation to
another.
16.5.4 Choosing the Technique to Fit
the Situation
There are often obvious circumstances that suggest
the most suitable technique for a given valuation
problem. For example, contingent valuation is the
only accepted method available for measuring existence values. We may rule out a benefits transfer
approach to provide information for a benefit-cost
analysis of a proposed policy decision if the possible cost of an error is very great relative to the po-