Ecological Economics 58 (2006) 304 – 317
www.elsevier.com/locate/ecolecon
ANALYSIS
Valuing the diversity of biodiversity
Mike Christie a,*, Nick Hanley b, John Warren a, Kevin Murphy c, Robert Wright b,
Tony Hyde d
a
c
Institute of Rural Sciences, University of Wales Aberystwyth, SY23 3AL, UK
b
Economics Department, University of Stirling, FK9 4LA UK
Department of Environmental and Evolutionary Biology, Glasgow University, G12 8QQ UK
d
Socio-Economic Research Services, Aberystwyth, SY23 3AH UK
Received 3 August 2004; received in revised form 1 June 2005; accepted 6 July 2005
Available online 4 October 2005
Abstract
Policy makers have responded to concerns over declining levels of biodiversity by introducing a range of policy measures
including agri-environment and wildlife management schemes. Costs for such measures are relatively easy to establish, but
benefits are less easily estimated. Economics can help guide the design of biodiversity policy by eliciting public preferences on
different attributes of biodiversity. However, this is complicated by the generally low level of awareness and understanding of
what biodiversity means on the part of the general public. In this paper we report research that applied the choice experiment
and contingent valuation methods to value the diversity of biological diversity. Focus groups were used to identify ecological
concepts of biodiversity that were important and relevant to the public, and to discover how best to describe these concepts in a
meaningful and understandable manner. A choice experiment examined a range of biodiversity attributes including familiarity
of species, species rarity, habitat, and ecosystem processes, while a contingent valuation study examined public willingness to
pay for biodiversity enhancements associated with agri-environmental and habitat re-creation policy. The key conclusions
drawn from the valuation studies were that the public has positive valuation preferences for most, but not all, aspects of
biodiversity, but that they appeared to be largely indifferent to how biodiversity protection was achieved. Finally, we also
investigate the extent to which valuation workshop approaches to data collection can overcome some of the possible
information problems associated with the valuation of complex goods. The key conclusion was that the additional opportunities
for information exchange and group discussion in the workshops helped to reduce the variability of value estimates.
D 2005 Elsevier B.V. All rights reserved.
Keywords: Biodiversity; Agri-environmental policy; Choice experiments; Contingent valuation; Valuation workshops
1. Introduction
* Corresponding author. Tel.: +44 1970 622217; fax: +44 1970
611264.
E-mail address: mec@aber.ac.uk (M. Christie).
0921-8009/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.ecolecon.2005.07.034
Society needs to make difficult decisions regarding
its use of biological resources. For example in terms
of habitat conservation, or changing how we manage
M. Christie et al. / Ecological Economics 58 (2006) 304–317
farmland through agri-environmental policy (Hanley
and Shogren, 2001). Environmental valuation techniques can provide useful evidence to support such
policies by quantifying the economic value associated
with the protection of biological resources. Pearce
(2001) argues that the measurement of the economic
value of biodiversity is a fundamental step in conserving this resource since dthe pressures to reduce biodiversity are so large that the chances that we will
introduce incentives [for the protection of biodiversity] without demonstrating the economic value of
biodiversity are much less than if we do engage in
valuationT. OECD (2001) also recognises the importance of measuring the economic value of biodiversity
and identifies a wide range of uses for such values,
including demonstrating the value of biodiversity, in
targeting biodiversity protection within scarce budgets, and in determining damages for loss of biodiversity in liability regimes.
More generally, the role of environmental valuation
methodologies in policy formulation is increasingly
being recognised by policy makers. For example, the
Convention of Biological Diversity’s Conference of
the Parties decision IV/10 acknowledges that
deconomic valuation of biodiversity and biological
resources is an important tool for well-targeted and
calibrated economic incentive measuresT and encourages Parties, Governments and relevant organisations to dtake into account economic, social, cultural
and ethical valuation in the development of relevant
incentive measuresT.
1.1. Valuing biodiversity: the challenge
However, what concerns us here is not whether one
should attempt to place economic values on changes
in biodiversity, but rather in what the particular difficulties are in doing so. These include incommensurate
values or lexicographic preference issues (Spash and
Hanley, 1995; Rekola, 2003) and–the issue we focus
on here–people’s limited understanding of complex
environmental goods (Hanley et al., 1996; Christie,
2001; Limburg et al., 2002).
Stated preference valuation methods require survey
respondents to make well-informed value judgements
on the environmental good under investigation. This
requires information on unfamiliar goods to be presented to respondents in a meaningful and understand-
305
able format. Herein lies the problem: many studies
have found that members of the general public have a
low awareness and poor understanding of the term
biodiversity, and that communicating relevant information within a stated preference study is difficult.
Furthermore, if one is unaware of the characteristics
of a good, then it is unlikely that one has well-developed preferences for it which can be uncovered in a
stated preference survey.
Various surveys have examined the publics’ understanding of the term dbiodiversityT. A recent UK survey found that only 26% of respondents had heard of
the term dbiodiversityT (DEFRA, 2002). Similar findings are also reported in Spash and Hanley (1995) and
Luthbert (2002). The lack of public understanding of
the term biodiversity will make the valuation exercise
difficult; however, people can learn during a survey,
and may have preferences for what biodiversity actually means, even if they are unaware of the term itself:
the DEFRA (2002) survey also found that 52% considered the protection of wildlife to be dvery
importantT, even though they did not know what
biodiversity itself meant.
A related complication is that biodiversity itself is
not uniquely defined by conservation biologists.
Scientists are in general agreement that the number
of species per unit of area provides a useful starting
point (Harper and Hawksworth, 1995; Whittaker,
1977). Although such a measure appears to be relatively straightforward, issues such as what constitutes
a species (Harper and Hawksworth, 1995; Claridge
and Boddy, 1994); and what size of area to count
species over complicate this measure (Whittaker,
1977). Even if these questions were resolved, ecologists recognise that some species, such as keystone
species, may be more important and/or make a
greater contribution to biodiversity than others (Wilson et al., 2003; Noss, 1990). A further complicating
factor relates to the extent to which the public is
capable of understanding these ecological concepts.
Ecologists also recognise that biodiversity may be
described and measured in terms of species diversity
within a community or habitat (Arts et al., 1990) and
in terms of the diversity of ecological functions (Steneck and Dethier, 1994; Herrera et al., 1997). Finally,
the public may have stronger preferences for certain
species that display charismatic features such as
beauty or speed, or be locally significant, even though
306
M. Christie et al. / Ecological Economics 58 (2006) 304–317
these features may not be considered ecologically
important (May, 1995).
The issues highlighted above indicate that research
that attempts to value changes in biodiversity using a
direct elicitation of public preferences will be challenging, since it requires us to identify appropriate language in which complex biodiversity concepts can be
meaningfully conveyed to members of the public in
ways which are consistent with underlying ecological
ideas on what biodiversity is.
This paper aims to identify problems surrounding
the economic valuation of dbiodiversityT. In particular, we report the results from a series of stated
preference studies on changes in biodiversity on
UK farmland. The studies include a contingent valuation study on three biodiversity enhancing policies
(agri-environment scheme, habitat re-creation and
protection of biodiversity loss associated with housing development) and a choice experiment that examines the value of biodiversity attributes (familiar
species of wildlife, rare unfamiliar species of wildlife,
habitats and ecosystem services). We also examine
through a series of valuation workshops the impact of
information deficit which typifies the knowledge
level of most members of the general public regarding biodiversity.
The paper is organised as follows. Section 2 presents a brief review of the current literature on valuing
biodiversity and identifies gaps in this literature. Our
study design is explained in Section 3, with results
presented in Section 4. A discussion concludes the
paper.
2. Previous literature
A general comment on much of the existing biodiversity valuation literature is that it mostly does not
value diversity itself, but rather focuses on individual
species and habitats (Pearce, 2001). In this section, we
review a number of key studies that have attempted to
measure the economic value of different elements of
biodiversity. In particular, we distinguish between
studies that have valued a biological resource (e.g. a
particular species, habitat area, or ecosystem function)
and those which have valued the biological diversity
of those resources (e.g. ecological concepts of biodiversity such as the rarity of a species).
2.1. Studies that value biological resources
There have been a large number of studies that
have valued particular species. Most of these studies
have been undertaken in the US and utilise stated
preference techniques, thus enabling both use and
passive-use values to be assessed. Nunes and van
den Bergh (2001) provide an extensive review of
valuation studies that have addressed both single
and multiple species. Valuations for single species
range from $5 to $126 per household per year, and
for multiple species range from $18 to $194. In the
UK, there have been a limited number of studies that
have valued both single and multiple species. For
example, Macmillan et al. (2002) estimated the value
of wild geese conservation in Scotland, while White
et al. (1997, 2001) examine the value associated with
the conservation of four UK mammals: otters, water
voles, red squirrels, and brown hare. Macmillan et al.
(2001) also takes a slightly different perspective by
valuing the reintroduction of two species (the beaver
and wolf) into native forests in Scotland.
Biological resources may also be considered in
terms of the diversity within natural habitats. Studies
have addressed the valuation of habitats from two
perspectives. One approach is to link the value of
biodiversity to the value of protecting natural areas
that have high levels of outdoor recreation or tourist
demand. A second approach to the valuation of natural areas involves the use of stated preference methods. UK examples of contingent valuation (CV)
studies that have valued habitats include: Garrod
and Willis (1994) who examined the willingness to
pay of members of the Northumberland Wildlife Trust
for a range of UK habitat types; Hanley and Craig
(1991) who valued upland heaths in Scotland’s flow
country; and Macmillan and Duff (1998) who examine the publics’ willingness to pay (WTP) to restore
native pinewood forests in Scotland.
Ecosystem functions and services describe a wide
range of life support systems including waste assimilation, flood control, soil and wind erosion prevention, and water quality maintenance (Fromm, 2000).
Many of these functions and services are complex and
it is likely that members of the public will possess a
poor understanding of these issues. The consequence
of this is that attempts to value ecosystem functions
and services will be difficult, particularly in methods
M. Christie et al. / Ecological Economics 58 (2006) 304–317
(such as the stated preference methods) where respondents are required to make a value judgement based
on the description of the good in question. Analysts
often use other techniques including averting behaviour, replacement costs, and production functions to
measure the indirect values of ecosystem functions.
2.2. Studies that value the diversity of biological
resources
Studies that have quantified genetic diversity have
predominantly measured direct use benefits of biological resources in terms of inputs to the production of
market goods such as new pharmaceutical and agricultural products. The majority of studies have based
valuations on market contracts and agreements for
bioprospecting by pharmaceutical industries (Simpson
et al., 1996; Rausser and Small, 2000). Ten Kate and
Laird (1999) provide an extensive review of such
bioprospecting agreements. Franks (1999) provides a
useful contribution on the value of plant genetic
resources for food and agriculture in the UK and
also the contribution of the UK’s agri-environmental
schemes to the conservation of these genetic
resources.
A number of valuation studies have attempted to
value biodiversity by explicitly stating to respondents
that the implementation of a conservation policy will
result in a change in the biodiversity of an area. For
example, Garrod and Willis (1997) estimated passiveuse values for biodiversity improvements that
increased the proportion of broad-leaved trees planted
and the area of open spaces in the forest in remote
upland coniferous forests in the UK. Willis et al.
(2003) extend this work to examine public values
for biodiversity across a range of UK woodland
types. Other studies have assessed public WTP to
prevent a decline in biodiversity. For example, Macmillan et al. (1996) measures public WTP to prevent
biodiversity loss associated with acid rain; while
Pouta et al. (2000) estimate the value of increasing
biodiversity protection in Finland through implementing the Natura 2000 programme.
White et al. (1997, 2001) examine the influence of
species characteristics on WTP. They conclude that
charismatic and flagship species such as the otter
attract significantly higher WTP values than less
charismatic species such as the brown hare. They
307
further suggest that species with a high charisma
status are likely to command higher WTP values
than less charismatic species that may be under a
relatively greater threat or of more biological significance in the ecosystem. In a meta-analysis of WTP
for a range of species, Loomis and White (1996) also
find that more charismatic species, such as marine
mammals and birds, attract higher WTP values than
other species.
The above review has demonstrated that from
those studies that have claimed to value biodiversity,
only a handful have actually examined the diversity
that exists within biological resources; most studies
have alternatively tended to simply value a particular
biological resource such as a species, habitat or ecosystem service. Furthermore, studies that have
attempted to value the diversity of biological
resources currently only provided limited information
on the value of the components of biological diversity.
Research effort has yet to provide a comprehensive
assessment of the value attached to the components of
biological diversity such as anthropocentric measures
(e.g. cuteness, charisma, and rarity) and ecological
measures (e.g. keystone species and flagship species).
It is this issue of valuing the ecological and anthropocentric diversity of biological resources that the
current research aims to address.
3. Study design
3.1. Developmental focus groups
The policy setting for this research is the development of policy on biodiversity conservation and
enhancement on farmland in England. The principal challenges in study design were to identify
which ecological and anthropocentric concepts of
biodiversity are needed to be communicated to the
general public, and thus form the focus of the
valuation exercise. We also needed to design effective ways of conveying the complex information
on biodiversity.
In a review of ecological literature (Christie et al.,
2004), we identified 11 different concepts that ecologists commonly use to describe and measure biodiversity. Clearly, it would be extremely difficult to
attempt to value all of these concepts within a single
308
M. Christie et al. / Ecological Economics 58 (2006) 304–317
economic valuation study. In an attempt to simplify, a
conceptual framework was drawn up to provide a
framework in which public understanding of biodiversity could be tested (Fig. 1). This framework is
split into sections according to which perspective we
take on the importance and meaning of biodiversity:
ecological or anthropocentric. Within each of these
headings, we identify different aspects of biodiversity
that need to be considered for inclusion. The final row
of the figure shows the biodiversity attributes that
were eventually selected for the experimental design
of the choice experiment. We now explain how these
were chosen.
A series of focus groups comprising members of the
general public were arranged. Discussions aimed to
identify the level of understanding that the public had
for each of the elements of the framework in the third
line of Fig. 1, and also to identify their views on the
importance of each element. The key issues identified
in the developmental focus groups included:
! Over half of the participants could not remember
having come across the term dbiodiversityT before.
Some of those who had indicated a familiarity with
the term dbiodiversityT were unable to provide a
clear or accurate definition of the concept.
! Participants indicated that they were familiar with
related terms including dspeciesT, dhabitatT and
decosystemT.
! Participants indicated that they were not familiar
with the majority of scientific concepts of biodiversity in Fig. 1 such as keystone species and
flagship species. On a more positive note, it was
also found that most participants of the focus
groups appeared to be capable of quickly picking
up a basic understanding of most biodiversity
concepts if these were explained in layman’s
terms.
The conclusion from this is that the survey would
need to employ alternative, non-scientific terminology
to meaningfully describe the ecological concepts associated with biodiversity. Based on this focus group
evidence, four attributes were identified as being
appropriate to describe the diversity of biodiversity
concepts to the public:
! Familiar species of wildlife. This attribute includes
charismatic, familiar (recognisable) and locally
symbolic species, and may be considered in
terms of both common and rare familiar species.
! Rare, unfamiliar species of wildlife. This attribute
focuses on those species that are currently rare or in
decline which are unlikely to be familiar to members of the public.
! Habitat. The protection of habitats and in particular the mix of species that reside within them was
considered to be an important component of biodiversity conservation. Of note in this category was
the fact that focus group participants were more
concerned about achieving a biodiversity outcome
(i.e. protecting the range of species within a habitat), rather than a focus on how this might be
achieved (e.g. by targeting policy towards the protection of ecologically significant species such as
keystone or umbrella species).
! Ecosystem processes. The public was also concerned with preserving the dhealthT of ecosystem
Fig. 1. Conceptual framework used in the experimental design.
M. Christie et al. / Ecological Economics 58 (2006) 304–317
processes. It was also considered useful to distinguish between ecosystem processes which have a
direct impact on humans and those which do not.
Another question that needed to be addressed
relates to which methodology is likely to be the
most suited to the valuation of biodiversity change.
In this study, we aim, if possible, to capture all
components of the total economic value associated
with biodiversity change. Stated preference methods
(including contingent valuation and choice experiments) appear to be the most flexible valuation
approach since they are capable of capturing both
use and passive-use values. A key objective of this
research was to measure the economic value of the
component attributes of biological diversity. It was
concluded that choice experiments would be the most
appropriate method to value these attributes (Bennett
and Blamey, 2001). In addition, we also wanted to
estimate values for three types of biodiversity changes
that were considered to be of particular relevance to
policy makers, namely: biodiversity enhancement
associated with agri-environmental schemes, biodiversity enhancements associated with the re-creation
of wildlife habitats, and biodiversity loss from farmland associated with development activities (e.g. house
building). Contingent valuation scenarios could be
designed to directly elicit the values of the three proposed policy programmes, and thus seems a neater,
more direct approach with regard to this second
research objective. Thus, this research involved the
use of both the choice experiment and contingent
valuation methods to value biodiversity. In-person
household interviews were undertaken with a random
sample of the population in two case study areas.
Cambridgeshire was chosen as a predominantly intensively arable area that supports low levels of biodiversity, while Northumberland was chosen to represent an
area with high levels of biodiversity and a lower
intensity of land use.
3.2. Design specifics
A key factor affecting the validity of stated preference studies relates to the success to which the
good under investigation can be meaningfully, accurately, and consistently presented to survey respondents. Although this can be a challenge in many
309
valuation studies, the very fact that only a small
proportion of the public have knowingly heard of
the term biodiversity before presents a significant
challenge to this research. In this study, the survey
instrument was required to present a lot of information
on biodiversity which is likely to be complex and new
to respondents. The majority of valuation studies tend
to describe the environmental good under investigation using verbal descriptions, perhaps supported by
some written script and/or pictorial images. Although
such an approach to presenting the good can be
successful with goods that are familiar to survey
respondents, evidence gathered in the developmental
focus groups indicated that such a standard approach
was unlikely to be suitable for presenting biodiversity
which was found to be unfamiliar and considered
complex. Feedback from the focus groups also indicated that the large volume of new information
required to be presented on biodiversity was found
to lead to both confusion and respondent fatigue. The
adoption of a more visual and interactive approach
was therefore considered to be more suitable.
For these reasons, a PowerPoint dslideshowT was
used to convey information to respondents. This has a
number of advantages in terms of using a range of
formats (pictures, audio tracks and text), which helps
minimise respondent fatigue, reduces variation in the
information presented, and maximise the effectiveness
with which information is conveyed. The PowerPoint
presentation introduced survey respondents to a simple definition of biodiversity; dbiodiversity . . . is the
scientific term used to describe the variety of wildlife
in the countrysideT. Slides 3 to 8 then introduced the
four attributes of biodiversity that had been identified
in the developmental focus groups: familiar species of
wildlife, rare (unfamiliar) species of wildlife, habitat,
and ecosystem processes. Each attribute was defined,
and alternative levels of biodiversity enhancements
associated with these attributes were introduced.
Within these descriptions, named examples of relevant species, habitats and ecosystem processes within
the study areas were provided and images presented.
These were included to help respondents attain a
clearer understanding of the various aspects of biodiversity being discussed. Following the presentation of
this information, respondents were provided with an
opportunity to discuss and clarify with the interviewer
any issues of outstanding confusion. In slides 9–12,
310
M. Christie et al. / Ecological Economics 58 (2006) 304–317
the case study area (Cambridgeshire or Northumberland) was then introduced. Details presented included
a description of the predominant land uses found
within the case study area, and the current levels of
biodiversity that exist in that area. Respondents were
then informed that human activities, such as farming
and development, are currently threatening overall
levels of biodiversity in the area and the consequences
of this on the four biodiversity attributes were outlined. Slides 13–18 informed respondents that the
government could introduce policies to help protect
and enhance biodiversity in the respective case study
areas. Policies described included agri-environmental
schemes and habitat re-creation schemes. Slides 14–
17 then outlined how such policies could be introduced to specifically enhance the four aspects of
biodiversity identified earlier. In each case, the potential improvements were described in terms of the
attribute levels used in the choice experiment. Respondents were then asked to think about which aspects of
biodiversity they would like to see being protected and
enhanced. Finally, at the end of the presentation
respondents were given a further opportunity to clarify
any issues of confusion/uncertainty regarding any
aspect of the presentation.
Feedback from respondents of a pilot survey indicated that the majority of respondents understood the
concepts presented. Respondents also indicated that
the presentation of more information (to try to
increase understanding) would lead to respondent
fatigue. The inclusion of opportunities for respondents
to clarify and discuss issues of confusion with the
interviewer was seen as a valuable option. The impact
of information provision was an issue that was further
explored in a series of valuation workshops, as
explained below.
3.2.1. The choice experiment
Following the PowerPoint presentation, respondents of both the household survey and valuation
workshops were asked to complete a choice experiment exercise. The choice experiment was introduced
as follows:
In the presentation you were provided with information on different aspects of biodiversity. You
were also informed that biodiversity within Cambridgeshire (Northumberland) is under threat. We
as a society have some options over how we
respond to the threats to biodiversity. We are therefore interested in your opinions on what action you
would most like to see taken.We are now going to
show you five alternative sets of policy designs that
could be used to enhance Cambridgeshire’s (Northumberland’s) biodiversity. In each set, you will be
asked to choose the design which you prefer.
An example of a choice task was then presented to
respondents and the choice task was explained. Once
the respondents had undertaken all five choice tasks,
they were asked to indicate the main reason that they
had for making the choices that they did. This was to
allow protest responses to be identified.
We have already explained how biodiversity attributes were selected for inclusion in the choice experiment (above). Each of these attributes was then
defined according to three levels of provision, including the status quo (i.e. ddo nothingT which would lead
to a continued decline in biodiversity in the study
area) and two levels of improvement/enhancement.
Table 1 provides a summary of the four biodiversity
attributes used in the choice experiment, along with
the three levels of provision of each attribute.
The payment vehicle used in the choice experiment
was an annual increase in taxation over the next 5
years. The reasons for using this payment vehicle
include the fact that biodiversity enhancement programmes are generally paid for through taxation and
that participants of the focus groups indicated that
taxation was their preferred payment option. Five payment levels of taxation were used in the choice experiment. SPSS dOrthoplanT was used to generate a
(34 51) fractional factorial experimental design,
which created 25 choice options. A blocking procedure
was then used to assign the options to 10 bundles of
five choice sets. Thus each choice experiment respondent was presented with a bundle of five choice tasks.
3.2.2. The contingent valuation
Three biodiversity conservation policy scenarios
were presented in the contingent valuation study.
These were:
n WTP for an agri-environmental scheme that incorporated conservation headlands, and reduced use of
pesticides and fertilisers (Cambridgeshire only).
311
M. Christie et al. / Ecological Economics 58 (2006) 304–317
Table 1
Summary of biodiversity attributes and levels used in the choice experiment
Familiar species of
wildlife
Rare, unfamiliar
species of wildlife
Habitat quality
Ecosystem process
Annual tax increase
Policy level 1
Policy level 2
Do nothing (biodiversity
degradation will continue)
Protect rare familiar
species from further
decline
Slow down the rate
of decline of rare,
unfamiliar species
Habitat restoration, e.g.
by better management
of existing habitats
Only ecosystem services
that have a direct impact
on humans, e.g. flood
defence, are restored
Protect both rare and
common familiar species
from further decline
Stop the decline and
ensure the recovery of
rare unfamiliar species
Habitat re-creation, e.g.
by creating new habitat
areas
All ecosystem services
are restored
Continued decline in the
populations of familiar
species
Continued decline in the
populations of rare,
unfamiliar species
Wildlife habitats will
continue to be degraded
and lost
Continued decline in the
functioning of ecosystem
processes
10
25
n WTP for habitat creation: involving seasonal flood
plains, reed beds and more natural river flows
(Cambridgeshire) and the creation of wet grassland
(Northumberland).
n WTP to protect farmland currently under agrienvironmental schemes from development in
the form of new houses (Cambridgeshire and
Northumberland).
In both case study areas, respondents of the household survey would receive only one of these three
scenarios. In all of the above cases, the levels of
biodiversity change were described in terms of the
biodiversity attributes used in the choice experiment.
Annual increases in taxation over the next 5 years
were again the chosen payment vehicle in the CV
study, and these were presented to respondents using
the payment card elicitation method.
3.2.3. The valuation workshops
Six valuation workshops (Macmillan et al., 2002)
were undertaken in Northumberland. The workshops
used the same survey instrument as the household
study, but the structure of the workshops allowed
much greater time for reflection on the information
provided, while participants were encouraged to discuss the issues involved with each other. Opportunities for questions to the moderator also existed.
Following these discussions, participants were asked
100
260
520
No increase in
your tax bill
to complete a further series of five choice experiment
tasks. The aim of this was to explore the impact of
further information provision and discussion on value
judgements.
4. Results
In the main household survey, 741 respondents
(343 in Cambridgeshire and 398 in Northumberland)
each undertook five choice experiment tasks and a
single contingent valuation scenario. In the valuation
workshops, 53 respondents (Northumberland only)
undertook five choice experiment tasks before the
discussion and five choice tasks after the discussion.
4.1. Choice experiment results
The data from the choice experiment method were
analysed using a conditional logit model (see Louviere et al., 2000 for a detailed description of this
method of analysis). Welfare estimates in the form of
implicit prices (IP) were derived from the conditional
logit model using the following formula
IP ¼
bAttribute
:
bM
ð3Þ
where b Attribute is the coefficient on the attribute of
interest and b M is the negative of the coefficient
312
M. Christie et al. / Ecological Economics 58 (2006) 304–317
directly-relevant services aloneT (o53.62) unexpectedly creates higher utility than dall ecosystem
servicesT (o42.21). Another unexpected result comes
from the rare, unfamiliar species attribute. Here,
although a move from continued decline to stopping
decline and densuring recoveryT substantially increases well-being (o115.15), a move to dslowing
down declineT is negatively valued ( o46.68).
For Northumberland, the same pattern is
repeated, except that the dall ecosystem servicesT
and dslow down decline of rare unfamiliar speciesT
attributes are not significant. This means that any
improvement in dhabitat restorationT, dhabitat recreationT, ‘protection of rare familiar species onlyT
and the dprotection of both rare and common familiar speciesT are positively and significantly valued,
as is an improvement in ddirectly-relevant ecosystem
servicesT—although not an improvement in dall ecosystem servicesT. This implies that the Northumberland group only cared about ecosystem services that
of the monetary variable, namely the annual tax
increase.
Table 2 shows results from the choice experiment
data for both Cambridgeshire and Northumberland.
The pseudo-R 2 value is higher for the latter sample,
and is very close to the 20% level suggested by
Louviere et al. (2000) as indicating a very good fit
in this kind of data. The Cambridgeshire model shows
significant estimates for all the biodiversity attributes.
In almost all cases, parameter signs are in accordance
with a priori expectations and a scale effect is present
in most cases. Improving familiar species from continued decline to either dprotecting rare familiar species onlyT or dprotecting both rare and common
familiar speciesT increases utility by o35.65 and
o93.49 annually for the next 5 years respectively.
Moving the habitat attribute from continued decline
to dhabitat restorationT (o34.40) or habitat re-creation
(o61.36) is also positively valued. Moving ecosystem
services from continued decline to a drecovery of
Table 2
Conditional logit models and implicit prices for household survey choice experiment
Attribute
Parameter
estimate
t-value
Cambridgeshire
FAMRARE
FAMBOTH
RARESLOW
RARERECOVER
HABRESTORE
HABCREATE
ECOHUMAN
ECOALL
PRICE
Pseudo-R 2
N (individuals)
0.126
0.331
0.165
0.408
0.122
0.217
0.19
0.15
0.004
14%
343
2.1*
5.2*
3.0*
5.7*
2.3*
3.5*
3.2*
2.2*
15.2*
35.65
93.49
46.68
115.13
34.40
61.36
53.62
42.21
17.19
18.03
15.88
21.22
15.32
17.52
16.97
19.23
1.95
58.15
77.80
73.53
4.37
27.02
20.35
4.51
69.34
128.82
15.55
156.72
64.42
95.69
86.88
79.90
Northumberland
FAMRARE
FAMBOTH
RARESLOW
RARERECOVER
HABRESTORE
HABCREATE
ECOHUMAN
ECOALL
PRICE
Pseudo-R 2
N (individuals)
0.309
0.334
0.08
0.645
0.243
0.253
0.359
0.064
0.003
19%
398
5.1*
5.2*
1.5
8.1*
4.7*
4.3*
5.9*
1.0
15.3*
90.59
97.71
n/a
189.05
71.15
74.00
105.22
n/a
19.24
18.47
n/a
25.28
16.29
17.51
17.7
n/a
52.87
61.50
n/a
139.50
39.22
39.68
70.52
n/a
128.30
133.91
n/a
238.59
103.07
108.31
139.91
n/a
* Indicates significance at p = 0.05.
Implicit price
(o per annum)
SE
95% lower
(o)
95% upper
(o)
M. Christie et al. / Ecological Economics 58 (2006) 304–317
seemed to directly impact on their well-being. The
Northumberland group also had a negative value for
dslowing down the decline of rare unfamiliar
speciesT, but this estimate is insignificant. The statistical equivalence of the parameter estimates of the
two models can be compared using a Likelihood
Ratio test. The probability value for this test is
b 0.01, indicating that the models are different. In
other words, the valuation of biodiversity attributes
varies significantly between the two case study
areas.
4.2. Comparison of household study and valuation
workshop
In Table 3, models are presented for the choice
exercises undertaken during the valuation workshops.
Participants made two sets of choices, one near the
outset, after receiving the same information as the
household survey participants (referred to as dBeforeT),
and one near the end, having had a chance to discuss
and reflect on the issues further (referred to as dAfterT).
Neither model fits very well due to the small sample
size, but we can note that the number of significant
variables increases from three to seven between the
two treatments, while the overall fit also improves. In
other words, a learning effect seems to be present.
Looking at the dAfterT model, we see that it compares
Table 3
Conditional logit models for valuation workshop choice experiment,
Northumberland
Attribute
FAMRARE
FAMBOTH
RARESLOW
RARERECOVER
HABRESTORE
HABCREATE
ECOHUMAN
ECOALL
TAX
A _ OPTA
A _ OPTB
2 ln L
p-value
Pseudo-R 2
N (individuals)
dBeforeT discussion
dAfterT discussion
Parameter
t-statistic
Parameter
t-statistic
0.172
0.257
0.028
0.166
0.093
0.323
0.386
0.116
0.004
0.823
0.894
417.4
b0.01
16.7%
53
1.1
1.6
0.2
0.8
0.7
2.0*
2.4*
0.6
6.2*
2.3
2.4
0.327
0.343
0.316
0.654
0.149
0.332
0.319
0.211
0.004
0.295
0.081
440.7
b0.01
18.7%
53
2.0*
2.0*
2.1*
3.0*
1.1
2.0*
2.0*
1.2
5.8*
0.8
0.2
* Indicates significance at p = 0.05.
313
quite well with the household survey CE results for
Northumberland (Table 3), with only the dslow down
decline of rare unfamiliar speciesT having a negative
sign, and with dall ecosystem servicesT still being
insignificant. The workshop choices also show habitat
restoration to have an insignificant effect on utility.
Implicit prices are also very similar to the main survey,
with a complete recovery of rare, unfamiliar species
having the highest welfare gain. Finally, we note that a
formal LR test shows that the parameters of the main
survey CE model for Northumberland are not significantly different for either the dBeforeT or dAfterT models from the valuation workshops. Thus, it would
appear that although the extra discussions in the workshops improved participant’s understanding of biodiversity concepts and thus allow them to state their
WTP more precisely, this extra level of knowledge
did not significantly influence their WTP for the biodiversity attributes. In this sense, the valuation workshops provide support for the main survey choice
experiment results.
4.3. Contingent valuation
Table 4 gives summary measures for the WTP bids
for the various biodiversity conservation scenarios in
Cambridgeshire and Northumberland, along with a
value for the dpooledT scenarios for each area. With
respect to the dpooledT results, about one-third of
respondents had a WTP of zero, in other words, did
not value these increases in biodiversity. Furthermore,
mean WTP is higher for Cambridgeshire respondents
(o58.87) than for those from Northumberland
(o42.47): this difference is statistically significant at
the 95% level. Median WTP is considerably less than
mean WTP in all cases, illustrating a common finding
in CV studies.
In Cambridgeshire, WTP is highest for agri-environmental schemes (o74.27), and lowest for preventing
development loss (o45.30). Habitat re-creation is
valued at o54.97. This is of general interest, since
the theory of loss aversion (Kahneman et al., 1991)
suggests that losses are often valued more than gains.
However, these changes are not symmetrical in our
case. What is more, these mean values are not statistically different from each other at 95% ( p = 0.11). In
Northumberland, WTP is higher for the habitat recreation scenario (o47.19) than the development loss
314
M. Christie et al. / Ecological Economics 58 (2006) 304–317
Table 4
WTP values for CV biodiversity scenarios in Cambridgeshire and Northumberland
95% Confidence
interval
95% Trimmed
mean
Median
% with
WTP= 0
Cambridgeshire
Agri-environment schemes
124
o74.27
o13.26
Habitat creation scheme
107
o54.97
o6.56
Development loss
110
o45.30
o7.82
All scenarios (bpooledQ)
341
o58.87
o5.84
Notes: F-test for difference in means: F = 2.2 and p = 0.11
o48.03 X o100.51
o41.96 X o67.98
o29.80 X o60.79
o47.38 X o70.36
o53.28
o48.42
o31.26
o42.84
o24.00
o24.00
o16.00
o20.00
29.8%
29.9%
37.3%
32.3%
Northumberland
Habitat creation Scheme
209
o47.49
o5.98
Development loss
186
o36.84
o5.07
All scenarios dpooledT
395
o42.47
o3.97
Notes: F-test for difference in means: F = 1.4 and p = 0.18
o35.70 X o59.27
o26.82 X o46.85
o34.67 X o50.27
o34.35
o25.29
o30.09
o12.00
o3.00
o10.00
27.8%
46.8%
35.9%
CV scenario
N
Mean
Standard
error
scenario (o36.84), but again this difference is not
significant ( p = 0.18).
The conclusions drawn from the CV data are that
people place positive values on increases in biodiversity. This value is higher in the Cambridgeshire sample than in the Northumberland sample. However, in
no cases does WTP differ across policy scenarios to a
significant degree. It thus appears that in this sample,
people care about increasing biodiversity, but not how
this is achieved.
5. Discussion
5.1. Do the public value biodiversity enhancements?
Two key questions which can be asked of these
data: is there evidence that the general public is willing to pay additional taxes to support biodiversity
conservation, and if so, then why?
With respect to the CE data, we are first interested
in whether respondents chose a biodiversity enhancement policy option (Option A or B) as opposed to
the ddo nothingT option. Some 85% of the choices
made by CE respondents were for choice option A or
B, demonstrating that the majority of respondents
were willing to pay for biodiversity enhancements.
Indeed, over half of the respondents (52.6%) stated
that they considered that the biodiversity improvements outlined in policy option A or B were dgood
value of my moneyT. Those respondents who consistently choose the status quo were recorded as gen-
uine zero bids and they accounted for 8% of responses:
3% stating that the biodiversity improvements were
not good use of their money, while 5% stated that they
already contribute to environmental causes. Protest
votes included dI do not think that increases in taxation
should be used to fund biodiversity improvementsT
(6.5%) and dThe costs of biodiversity improvement
should be paid for by those who degrade biodiversity’
(14.2%). Eighteen percent of the respondents stated
other reasons for their choices.
Equivalent analysis of the CV data indicates that
around two-thirds of respondents were WTP towards
the biodiversity enhancing projects. Of the one-third
respondents stating that they were not willing to pay
towards a biodiversity enhancing policy, 43.3% were
genuine zero bids and 38.4% were protest votes.
Thus, it would appear that the general public is
willing to pay additional taxes to protect and enhance
biodiversity.
5.2. What aspects of biodiversity do the public value
most?
Another question our research enables us to
address is bwhat aspects of biodiversity protection
policy do the public value most?Q Examining the
implicit prices from the choice experiment (Table 2)
provides some answers. Familiar species attained
positive and significant implicit prices. In Cambridgeshire, scale effects were evident in that the implicit
price for the protection of both rare and common
familiar species (o93.49) was significantly higher
M. Christie et al. / Ecological Economics 58 (2006) 304–317
than the protection of only the rare familiar species
(o35.65). This was not, however, the case in the
Northumberland sample, where the two levels of protection had similar implicit prices (o90.59 and o97.71
respectively for the protection of rare only and rare
and common familiar species). In conclusion, evidence from the choice experiment suggests that the
public does support policies that target rare familiar
species of wildlife, but the evidence is less clear for
common familiar species.
The second attribute addressed in the choice
experiment related to rare unfamiliar species of wildlife. Two levels of provision were addressed. The first
aimed to dslow down the rate of the declineT. The
second level aimed to dstop decline and ensure
recoveryT. The findings for the dslow downT attribute
level were interesting since it was found to be negative in the Cambridgeshire sample (indicating that
negative utility would be gained from a slow down
in the decline of the population of rare unfamiliar
species—which was not predicted), while the attribute
level was not significant in the Northumberland CE
model. The implications of this finding was that it
appears that the public is unwilling to support policies
that simply delay the time it takes for such species to
become locally extinct. This conclusion was further
emphasised by the fact that highest implicit prices
were attained from the dstop decline and ensure
recoveryT attribute level. Thus, the policy implication
of these findings is that the public appears to only
support policies that aim to achieve recovery of the
populations of rare unfamiliar species, rather than
those that simply attempt to slow down decline in
population numbers. A further implication of these
findings relates to the fact that survey respondents
were told that they were unlikely to ever see these
rare, unfamiliar species. Thus, these values can be
considered to represent passive-use values. Finally,
these results provide support for policies, such as
species Biodiversity Action Plans, which specifically
target rare, unfamiliar species.
The habitat attribute was included to assess
whether the public valued dthe restoration of existing
habitatsT or dthe re-creation of new habitatsT on farmland. Both attribute levels were found to be positive
and significant in the two case study locations. In
Cambridgeshire, the value for habitat restoration
(o35.65) was half that for habitat re-creation
315
(o61.36), while similar values were attained for both
levels in Northumberland (o71.15 and o74.01 respectively). Although the reason for this difference is
unclear, it is suggested that the Cambridgeshire
sample may have considered that there were very
few existing habitats within Cambridgeshire which
would benefit from restoration. However, there
was evidence that the public would support policies
that aimed to protect and enhance habitats, although
the value of the implicit prices were found to be
slightly lower than those found for the two species
attributes.
Finally, the ecosystem services attribute was
included to assess whether the public valued ecosystems that donly had a direct impact on humansT and
dall ecosystem services include those which did not
directly affect humansT. The ecosystems services that
had direct impacts on humans were found to be
positively and significantly valued. However, the
dall ecosystemsT attribute level was not significant
in the Northumberland model and was lower than
the dhuman impacts onlyT attribute level in the Cambridge sample. It would thus appear that survey
respondents dcaredT about ecosystem functions that
affect humans, but were less interested in the other
ecosystem services.
Based on the above analysis, it would appear that
the public does value most, but not all, of the biodiversity attributes included in our experimental design.
Although there is some evidence that the public
appears to be able to distinguish between alternative
attributes, it should be noted that the differences in the
values are generally not statistically different. Unfortunately, the reason for the lack of statistical significance is unclear; however, we postulate the following
possibilities. First, it may be that respondents simply
have similar WTP values for the different biodiversity
attributes. Alternatively, it may be that respondents
were indifferent to what aspects of biodiversity are
protected, but rather, as one respondent put it dall I’m
concerned about is that we protect our wildlife . . . I’m
not really concerned how this is achieved and to tell
the truth I don’t really know how this might best be
achieved . . . Surely it is best to let the scientists decide
what to doT. Differentiating between these two explanations has important policy implications. The first
explanation suggests that the public does have identifiable preferences for biodiversity attributes (all be it
316
M. Christie et al. / Ecological Economics 58 (2006) 304–317
similar preferences) and therefore from a consumer
sovereignty perspective, these preferences should be
taken into account in policy formulation (Hanley and
Shogren, 2001). In the case of the second explanation,
it would appear that the public does not have specific
and identifiable preferences of biodiversity attributes,
and therefore policy decisions on biodiversity protection could now be made without further reference to
public opinion. A final explanation for the lack of
significant differences between the value of biodiversity attributes may be that respondent’s limited knowledge of the individual biodiversity attributes may lead
to a high level of variability in their choices; indeed,
the findings from the workshop indicate that the
opportunity to further discuss the biodiversity attributes reduced the standard errors. Unfortunately,
insufficient data was collected in this research to
clearly distinguish between these explanations and
we suggest that further research be undertaken.
6. Conclusions
This study also stands out in that it is one of the
few studies that attempt to value the diversity of
biodiversity. Thus, rather than simply estimating the
value of a biological resource such as a particular
species or habitat, this research explores in detail
values for the ecological and anthropocentric concepts
that can be used to define and describe the diversity
that exists within biological resources.
Policy makers may benefit from information on the
economic value of different actions aimed at biodiversity protection, but also on which aspects of biodiversity are most valued by taxpayers. Stated
preference methods can provide both types of value
estimates, but implementing these methods is difficult
in this particular case since the general public has a
rather low level of understanding of what biodiversity
is and why it matters. In this study we make use of a
novel way of conveying information to respondents,
information which is consistent with ecological understanding of what aspects of biodiversity might be
considered. We then use choice experiments to estimate the relative values people place on these attributes, and contingent valuation to look at the value of
specific policy programmes. The conclusions drawn
from this are that the public has positive values for
biodiversity, but may be indifferent as to how biodiversity is actually protected.
How policy makers might choose to use such
information is something we have not addressed
here. One option is to use economics to set overall
budgets for biodiversity, but ask ecologists to determine how this money is targeted. Another option is to
use the kind of evidence presented above to use more
economic information in this targeting. But economists would argue that, in a world of scarce resources
and conflicting demands, some information on public
preferences for biodiversity conservation is better than
no information if society wishes to make sensible and
politically-inclusive choices.
Acknowledgements
We thank the Department of the Environment,
Food and Rural Affairs for funding this research,
and members of the steering committee for many
useful comments.
References
Arts, G.H.P., Roelofs, J.G.M., de Lyon, M.J.H., 1990. Differential
tolerances among softwater macrophyte species to acidification.
Canadian Journal of Botany 68, 2127 – 2134.
Bennett, J., Blamey, R., 2001. The Choice Modelling Approach to
Environmental Valuation. Edward Elgar, Cheltenham.
Christie, M., 2001. A comparison of alternative contingent valuation
elicitation treatments for the evaluation of complex environmental policy. Journal of Environmental Management 62, 255 – 269.
Christie, M., Warren, J., Hanley, N., Murphy, K., Wright, R., 2004.
Developing Measures for Valuing Changes in Biodiversity.
DEFRA, London. Report to.
Claridge, M.F., Boddy, 1994. Species recognition systems in insects
and fungi. In: Hawksworth, D.I. (Ed.), The Identification and
Characterisations of Pest Organisms. CAB International, Wallingford, pp. 261 – 274.
DEFRA, 2002. Survey of Public Attitudes to Quality of Life and to
the Environment—2001. DEFRA, London.
Franks, J.R., 1999. In situ conservation of plant genetic resources
for food and agriculture: a UK perspective. Land Use Policy 16
(2), 81 – 91.
Fromm, O., 2000. Ecological structure and functions of biodiversity
as elements of its total economic value. Environmental and
Resource Economics 16 (3), 303 – 328.
Garrod, G.D., Willis, N.D., 1994. Valuing biodiversity and nature
conservation at a local level. Biodiversity and Conservation 3,
555 – 565.
M. Christie et al. / Ecological Economics 58 (2006) 304–317
Garrod, G.D., Willis, K.G., 1997. The non-use benefits of enhancing forest biodiversity: a contingent ranking study. Ecological
Economics 21, 45 – 61.
Hanley, N., Craig, S., 1991. Wilderness development decisions and
the Krutilla–Fisher model: the case of Scotland’s dlow countryT.
Ecological Economics 4, 145 – 164.
Hanley, N., Shogren, S., 2001. Awkward choices: economics and
nature conservation. In: Bromley, D., Paavola, J. (Eds.), Economics, Ethics and Nature Conservation. Blackwells.
Hanley, N., Spash, S., Walker, L., 1996. Problems in valuing the
benefits of biodiversity conservation. Environmental and
Resource Economics 5, 249 – 272.
Harper, J.L., Hawksworth, D.L., 1995. Preface. In: Hawksworth,
D.L. (Ed.), Biodiversity: Measurement and Estimation. Chapman and Hall, London, pp. 5 – 12.
Herrera, R.A., Ulloa, D.R., Valdés-Lafont, O., Priego, A.G., Valdés,
A.R., 1997. Ecotechnologies for the sustainable management of
tropical forest diversity. Nature and Resources 33, 2 – 17.
Kahneman, D., Knetsch, J.L., Thaler, R.H., 1991. The endowment
effect, loss aversion and status quo bias. Journal of Economic
Perspectives 5, 193 – 206.
Limburg, K.E., O’Neill, R.V., Costanza, R., Farber, S., 2002.
Complex systems and valuation. Ecological Economics 41
(3), 409 – 420.
Loomis, J.B., White, D.S., 1996. Economic benefits of rare and
endangered species: summary and meta-analysis. Ecological
Economics 18, 197 – 206.
Louviere, J.J., Hensher, D.A., Swait, J.D., 2000. Stated Choice
Methods: Analysis and Application. Cambridge University
Press, Cambridge.
Luthbert, M., 2002. Survey of the public’s understanding of and
attitudes towards biodiversity. Unpublished Report: SNH.
Macmillan, D.C., Duff, E.I., 1998. The non-market benefits and
costs of native woodland restoration. Forestry 71 (3), 247 – 259.
Macmillan, D.C., Hanley, N., Buckland, S., 1996. Contingent valuation of uncertain environmental gains. Scottish Journal of Political Economy 43 (5), 519 – 533.
Macmillan, D.C., Duff, E.I., Elston, D., 2001. Modelling nonmarket environmental costs and benefits of biodiversity projects
using contingent valuation data. Environmental and Resource
Economics 18 (4), 340 – 391.
Macmillan, D.C., Philip, L., Hanley, N., Alvarez-Farizo, B., 2002.
Valuing the non-market benefits of wild goose conservation: a
comparison of interview and group-based approaches. Ecological Economics 43 (1), 49 – 59.
May, R.M., 1995. Conceptual aspects of the quantification of the
extent of biological diversity. In: Hawksworth, D.L. (Ed.),
Biodiversity: Measurement and Estimation. Chapman and
Hall, London, pp. 13 – 20.
317
Noss, R.F., 1990. Can we maintain biological and ecological integrity. Conservation Biology 4, 241 – 243.
Nunes, P.A.L.D., van den Bergh, J.C.J.M., 2001. Economic valuation of biodiversity: sense or nonsense? Ecological Economics
39, 203 – 222.
OECD, 2001. Valuation of Biodiversity Benefits: Selected Studies.
OECD, Paris.
Pearce, D., 2001. Valuing biological diversity: issues and overview.
In: OECD, D. (Ed.), Valuation of Biodiversity Benefits: Selected
Studies. OECD, Paris.
Pouta, E., Rekola, M., Kuuluvainen, J., Tahvonen, O., Li, C.-Z.,
2000. Contingent valuation of the Natura 2000 programme in
Finland. Forestry 73 (2), 119 – 128.
Rausser, G.C., Small, A.A., 2000. Valuing research leads bioprospecting and the conservation of genetic resources. Journal of
Political Economy 108 (1), 173 – 206.
Rekola,, 2003. Lexicographic preferences in contingent valuation: a
theoretical framework with illustrations. Land Economics 79
(2), 277 – 291.
Simpson, R.D., Sedjo, R.A., Reid, J.W., 1996. Valuing biodiversity
for use in pharmaceutical research. Journal of Political Economy
104 (1), 163 – 185.
Spash, C.L., Hanley, N., 1995. Preferences, information and biodiversity preservation. Ecological Economics 12, 191 – 208.
Steneck, R.S., Dethier, M.N., 1994. A functional group approach
to the structure of algal-dominated communities. Oikos 69,
476 – 498.
ten Kate, Laird, 1999. The Commercial Use of Biodiversity: Access
to Genetic Resources and Benefit-Sharing. Earthscan Publications, London.
White, P.C.L., Gregory, K.W., Lindley, P.J., Richards, G., 1997.
Economic values of threatened mammals in Britain: a case study
of the otter Lutra lutra and the water vole Arvicola terrestris.
Biological Conservation 82 (3), 345 – 354.
White, P.C.L., Bennett, A.C., Hayes, E.J.V., 2001. The use of willingness-to-pay approaches in mammal conservation. Mammal
Review 31 (2), 151 – 167.
Whittaker, R.H., 1977. Evolution of species diversity in land communities. Evolutionary Biology 10, 1 – 67.
Willis, K., Garrod, G., Scarpa, R., Powe, N., Lovett, A., Bateman,
I., Hanley, N., Macmillan, D., 2003. The Social and Environmental Benefits of Forests in Great Britain. Forestry Commission, Edinburgh.
Wilson, W.L., Abernethy, V.J., Murphy, K.J., Adam, A.,
McCracken, D.I., Downie, I.S., Foster, G.N., Furness, R.W.,
Waterhouse, A., Ribera, I., 2003. Prediction of plant diversity
response to land use change on Scottish agricultural land.
Agriculture, Ecosystems & Environment 94, 249 – 263.