Content uploaded by David Hoyos
Author content
All content in this area was uploaded by David Hoyos on Nov 19, 2014
Content may be subject to copyright.
Environ Resource Econ (2012) 51:119–140
DOI 10.1007/s10640-011-9491-9
Willingness to Pay for Ancillary Benefits of Climate
Change Mitigation
Alberto Longo · David Hoyos · Anil Markandya
Accepted: 31 May 2011 / Published online: 18 June 2011
© Springer Science+Business Media B.V. 2011
Abstract Assessing the Willingness to Pay (WTP) of the general public for climate change
mitigation programmes enables governments to understand how much taxpayers are willing
to support the implementation of such programs. This paper contributes to the literature on
the WTP for climate change mitigation programmes by investigating, in addition to global
benefits, the ancillary benefits of climate change mitigation. It does so by considering local
and personal benefits arising from climate change policies. The Contingent Valuation Method
is used to elicit the WTP for ancillary and global benefits of climate mitigation policies in the
Basque Country, Spain. Results show that WTP estimates are 53–73% higher when ancillary
benefits are considered.
Keywords Willingness to Pay · Contingent valuation · Ancillary benefits · Climate change
Mathematics Subject Classification (2000) Q51
1 Introduction
One of the main challenges that societies and economies will be facing in the coming years
is adapting to and mitigating climate change. The threat imposed by this phenomenon is
A. Longo (
B
)
School of Biological Sciences, Queen’s University Belfast—UKCRC Centre of Excellence
for Public Health (NI), Gibson Institute for Land, Food and the Environment, MBC Lisburn Road,
Belfast BT9 7BL, UK
e-mail: a.longo@qub.ac.uk
D. Hoyos
Department of Applied Economics III (Econometrics and Statistics), University of the Basque Country
(UPV/EHU), Lehendakari Aguirre, 83, 48015 Bilbao, Spain
e-mail: david.hoyos@ehu.es
A. Markandya
BC3. Basque Centre for Climate Change, Gran Vía, 35-2, 48009 Bilbao, Spain
e-mail: anil.markandya@bc3research.org
123
120 A. Longo et al.
becoming more prevalent due to an increase in carbon intensive production and consump-
tion patterns in both developed and developing countries. The Intergovernmental Panel on
Climate Change (IPCC) has reiterated in its latest report that societies should no longer
delay actions to curb global emissions of greenhouse gases (GHG) if they want to avoid
the catastrophic consequences of climate change (IPCC 2007). Public institutions around
the world are designing and implementing plans to mitigate the causes and to adapt to the
already existing impacts o f climate change. Both options come at a cost to society, as their
implementation requires costly technological and behavioural changes. As a consequence,
even though in principle the public may support climate change mitigation programs, it is
important to assess the extent of the support as households will eventually be presented with
a climate change packages bill, in terms of either higher prices, or taxes, or a combination
of both. Therefore, estimating the willingness to pay (WTP) of the current generation for
mitigating the effects of climate change is useful to determine the mitigation targets that
governments should undertake with the support of taxpayers (Layton and Brown 2000).
The public authorities of the Basque Country, Spain, have recently agreed to reduce by
16% current GHG emissions levels compared to 1990 levels through the implementation
of the “Basque Plan to Combat Climate Change 2008–2012” (BPCCC) (BG 2008). This
Plan would allow the Basque Country to be in line with the Spanish targets implied by the
Kyoto Protocol. The BPCCC states that, without further action, climate change would result
in the following outcome for the region by the end of the twentyfirst century: a reduction in
annual rainfall of between 15 and 20%, with wetter winters and d rier summers; maximum
temperatures that could rise by between 1.5 and 3.5
◦
C and minimum temperatures could
rise by between 1 and 3
◦
C; greater evapo-transpiration and increased risk of forest fires;
significant rises in sea levels and wave power leading to a receding coastline in beach areas
by between 11 and 3 m; increasing risk of flooding in estuaries. Measures identified by the
BPCCC to curb GHG emissions in the Basque Country entail, among others: increasing
the use of renewable energy sources and reducing the use of fossil fuels to produce energy,
adopting more energy efficient technologies, improving public transport, modernising waste
management by improving recycling schemes, as well as managing agricultural areas and
forests to reduce GHG emissions.
Economists have mainly used two approaches for evaluating climate change mitigation
policies. With the first approach, scientific aspects of climate change and economic agents’
behaviour are combined in macroeconomic models for estimating the social benefits and costs
of implementing mitigation policies (IPCC 2001; Dowlatabadi and Morgan 1993; Maddison
1995; Nordhaus 1994). Although they are able to include market and some non-market goods,
macroeconomic models rely on detailed and costly data on the world economy to infer the
benefits of climate change mitigation measures. Whilst these models are useful to assess
policies effects at a macro level, they fall short on explaining the behaviour of agents at the
micro level. To complement macroeconomic models, a second approach based on microeco-
nomic models using Stated Preference (SP) studies may be used to directly elicit individuals’
Willingness To Pay (WTP) for implementing climate change mitigation measures and assess
at a micro level the effects of future policies. In this paper, we use the Contingent Valuation
Method (CVM), an SP method, which is able to measure the total economic value, including
non-use values, to assess the benefits of climate change mitigation policies.
Recent studies using SP methods to assess the WTP of the general population for the total
economic value of climate change mitigation policies include Berk and Fovell (1999), Roe
et al. (2001), Berrens et al. (2004), Li et al. (2004), Nomura and Akai (2004), Li et al. (2005),
Hidano and Kato (2007), Longo et al. (2008)andTseng et al. (2009). All of these studies
have focused on the global benefits of GHG emissions reductions, and none of them has
123
Willingness to Pay 121
looked at ancillary benefits induced by climate change policies. Ancillary benefits of climate
change programs can be defined as positive externalities benefiting a local community, such
as reduction in local air pollution. Previous research has shown that ancillary benefits may be
substantial, as they have been estimated to lie in the range of a multiple of primary benefits
of between 0.98 and 6.93 (Pearce 2000). The scarce economic literature on ancillary benefits
has been mainly produced using macroeconomic models (Alfsen et al. 1992; Complainville
and Martins 1994) and through the valuation of the health benefits using the existing liter-
ature on use benefit estimation (OECD 2000).
1
This paper contributes to the literature on
ancillary benefits of climate change policies by using SP methods. Using the CVM, we elicit
the WTP for ancillary and global benefits of the BPCCC of a sample of 1,000 households
in the Basque Country to assess to what extent proposed programs to curb GHG emissions
have the support of the local population, and how this support depends on the local ancillary
benefits directly enjoyed by the affected population.
The paper is structured as follows. Section 2 reviews the literature on ancillary benefits of
climate change mitigation policies. Section 3 describes the survey instrument and its admin-
istration. Section 4 addresses the economic and econometric models used to answer our
research questions. Section 5 reports and discusses the empirical results and Sect. 6 provides
some concluding remarks and future research needs.
2 Ancillary Benefits of Climate Change
Ancillary benefits are sometimes known as secondary benefits, co-benefits or spillover ben-
efits, mainly depending on the relative emphasis given to their relation with the primary
effects. An example of ancillary benefits of climate policy is the positive health effects of air
pollution reduction accompanying a GHG emissions reduction. Some authors consider also
ancillary benefits the altruistic benefit or ‘warm glow’ that industrialised countries may enjoy
from helping more vulnerable and less responsible developing countries (Rübbelke 2002).
Other forms of ancillary benefits may be considered the benefits associated with employment
and technological effects in investments in green technologies or afforestation programmes
aiming to reduce CO
2
concentrations to improve the habitat for endangered species.
Primary and ancillary benefits of climate policies can be distinguished on the grounds
of three features: ‘publicness’, given that primary benefits are public goods while ancillary
benefits contain characteristics of public and private goods; delay, given that primary benefits
arise in the long run while ancillary benefits can be enjoyed in the short run; and the required
scientific knowledge, because the uncertainties around primary benefits exceed those around
the assessment of ancillary benefits (Markandya and Rübbelke 2004).
The literature on ancillary benefits has shown that they are substantial. Positive health
effects derived from GHG emissions reductions are usually considered the main category
of ancillary benefits (Ayres and Walter 1991; Heintz and Tol 1996; Olsthoorn et al. 1999).
1
Most studies, rather than looking at the direct value of the ancillary benefits of reducing GHGs, have esti-
mated the co-benefits in terms of US$ per ton of CO
2
abated. Earlier studies cited in the Third Assessment
Report of the IPCC (2001) and in the 2000 OECD review referred to above indicate a wide range of health
co-benefits, depending on country, assumptions of baselines, which pollutants were included, the dose-response
functions used and the valuation attached to different health end points, particularly premature mortality. The
range (in 2,000 prices) was from US$0.6 to US$145 per ton of CO
2
, with most under US$30. Since marginal
abatement costs from the power sector are estimated to be in the range of US$20–40 per ton of CO
2
,suchben-
efits could make a big difference in deciding which options to select for reducing GHGs. More recent studies
(see van den Bergh 2010) have provided some explanation for the wide range but they have not resulted in
really narrowing it. Hence an SP approach such as that adopted here may be a useful alternative to take.
123
122 A. Longo et al.
Air quality improvements reduce local mortality rates from air pollution and help mitigating
other negative effects, like surface corrosion, weathering of materials, impaired visibility and
vegetation acidic depositions. In the transport sector, ancillary benefits may include not only
the reduction of air pollution but also other social disturbances like noise, congestion and
road surface damage (Barker et al. 1993). A comprehensive overview of empirical studies
on ancillary benefits may be found in Rübbelke (2002), Burtraw et al. (2003), and Pittel and
Rübbelke (2008). These studies are heterogeneous from a methodological and geographical
perspective, which further complicates a comparison of their results. The majority of stud-
ies employ macroeconomic models applied to individual countries. The estimates found in
Western Europe usually exceed those found in studies from the USA. This difference has
been attributed to demographic and geographical differences, as well as different aggregation
levels. Interestingly, the consideration of ancillary benefits of climate mitigation in transi-
tion and developing economies has shown that they may be even higher than those obtained
for industrialised economies (Pittel and Rübbelke 2008; OECD 2000. Comparing results is
extremely difficult because of different ancillary benefits considered (air pollution, trans-
port activity, etc.), different baseline situations and mitigation reductions or even absence of
monetary estimations.
The main contribution of this paper to the scarce literature on ancillary benefits of cli-
mate change mitigation policies is in using an SP approach instead of the most common
macroeconomic models. We use respondents’ own perception of the local importance of
climate change and climate change programs in addition to a split sampling approach, where
half respondents are given details of the local effects of climate change and climate change
programs, to assess the value of ancillary benefits. In the following section, we describe the
survey instrument that we used to assess these ancillary benefits.
3 The Contingent Valuation Study
For this study, we administered a contingent valuation (CV) questionnaire to a representative
sample of the population of the Basque Country. Each respondent was taken to represent
his or her household. Two sets of focus groups were conducted that helped us in crafting a
questionnaire that was administered firstly to a sample of 100 households in May 2008. A
second pre-test of 100 respondents, conducted in June 2008, confirmed that the changes we
had made after the first wave of respondents were appropriate and no further indications for
any additional change had emerged. The final survey was administered in June and July 2008
when 1,000 completed interviews were collected. Both pre-tests and the final sample surveys
were conducted by professionally trained interviewees through in-person interviews at the
respondent’s home. The sample was selected to represent the population of households of
the Basque Country and was stratified in terms of age, gender and geographical distribution.
The CV questionnaire aimed at assessing to what extent citizens in the Basque Country
support climate change mitigation policies. In the questionnaire, each respondent faced three
dichotomous choice single-bounded WTP questions for implementing three programmes to
cut GHG emissions in the Basque Country: (1) by 4% compared to the current emissions
levels through an increase in the production of renewable electricity; (2) by 0.5% compared
to current emissions levels through the implementation of energy efficient measures in the
residential sector; (3) by 16% compared to the current emissions levels through the applica-
tion of the BPCCC that incorporates the previous two measures and a set of other measures
to reduce GHG emissions. For each programme, respondents were asked to state whether
they were willing to pay a proposed tax to fund the measure. Figure 1 provides a graphic
123
Willingness to Pay 123
Emissions reduction by means of energy efficiency measures
0%
2%
4%
6%
8%
10%
12%
14%
16%
Emissions reduction b y means of electricity production from renewable
sources
0%
2%
4%
6%
8%
10%
12%
14%
16%
Emissions reduction with this policy
Emissions reduction by means of a new tax on climate change
0%
2%
4%
6%
8%
10%
12%
14%
16%
Emissions reduction with this policy Emission reduction to comply with Kyoto targets
Emission reduction to comply with Kyoto targets
Emission reduction to comply with Kyoto targets
Emissions reduction with this policy
Fig. 1 Visual representation of greenhouse gases reductions shown in the questionnaire
123
124 A. Longo et al.
representation shown to respondents of the benefits in terms of GHG emissions reductions
entailed by the three programs.
As the questionnaire focused on eliciting households’ WTP for implementing the three
programmes to combat climate change, the survey instrument followed the standard struc-
ture of a CV study (Mitchell and Carson 1989). The questionnaire opened with “warm-up”
questions aimed at making the respondents comfortable with answering easy and general
questions also intended at understanding the relative importance to the respondents of cli-
mate change compared to other priorities, such as employment, education, public health, etc.
In the second part, respondents were queried about their knowledge of climate change and
were given information about the causes and the threat of climate change to elicit respon-
dents’ attitudes towards it. In the third part respondents were asked the dichotomous choice
CV questions, followed by debriefing questions to explain their answers to the WTP ques-
tions. In the fourth part, respondents were asked a set of attitudinal questions about energy,
such as their knowledge about the sources of energy used in the Basque Country, respondents’
energy expenditures and their transportation habits. The final section collected some socio-
economic characteristics of the respondents, including their age, civil status, occupation,
level of education, and income.
Following the standard CV practice, the payment vehicle used was a tax payable to the
Basque Government (Bateman et al. 2002; Carson and Hanemann 2005; Mitchell and Carson
1989). The climate mitigation programmes and payment vehicle were found to be both credi-
ble and understandable by focus groups participants. Following the literature on multiple CV
questions, the elicitation format used was single bounded dichotomous choice CV questions
with the order of the WTP for the three p rograms being randomized across respondents to
minimize ordering effects (Bannon et al. 2007; Hoehn and Loomis 1993; Park and Loomis
1996; Payne et al. 2000; Poe et al. 1997). Although this approach allows collecting high
quality data and minimizing protest answers, possible correlation among responses may
complicate comparisons of WTP estimates across scenarios (Poe et al. 1997), an issue that
we address in the econometric modelling approach (see below).
As standard use in dichotomous choice CV questions, different respondents were pre-
sented with different tax values. The set of six different bid amounts
2
ranged from e10 to
e180 for the tax for promoting renewable electricity production, from e5toe150 for the tax
for promoting energy efficient measures and from e20 to e350 for the tax for implementing
the whole BPCCC, as shown in Table 1. People were randomly assigned to one of six possible
sets of tax amounts.
The CV questions were followed by a set of debriefing questions to identify “protest”
respondents, i.e. respondents that state that they are not willing to pay the amount requested
but they may actually have a positive WTP (Bateman et al. 2002). We flagged as protest
respondents those who were not willing to pay because: (1) they complained that companies
are the major causes of climate change and therefore they should pay for it; (2) they sug-
gested that the government should pay for climate change, not the citizens; (3) they found
the proposed policies unrealistic; or (4) they felt that climate change was a global problem
and that not only Basques should pay for it.
2
The pretest bid vectors were the same as those used for the final questionnaire. The results from the pretest
showed that the percentage of ‘yes’ answers to the WTP questions decreased with the bid amount, and that
the bid vectors well captured the distribution of WTP. At the lowest bids the majority of respondents was
willing to pay the proposed bids, and as the bid amount increased the respondents willing to pay the proposed
amounts decreased to 12% for the BPCCC, 18% for the renewable energy program and 6% for the renewable
electricity program.
123
Willingness to Pay 125
Table 1 Bid values and
percentage of respondents
willing to pay the proposed bids
BPCCC Renewable electricity Energy efficiency
Bid % yes Bid % yes Bid % yes
20 66.44 10 82.5581.88
40 67.91 20 73.13 10 73.88
80 66.17 40 84.96 30 84.96
150 68.81 80 68.81 60 67.89
250 38.83 120 58.25 100 54.37
350 49.09 180 48.18 150 45.45
A central aspect of this study was to analyse the ancillary benefits of reducing GHG
emissions. We address the WTP for ancillary benefits of climate change in two ways: (1) by
varying the level of information across respondents for the damages of climate change and
the benefits of climate change mitigation programs; (2) and by investigating in the econo-
metric analysis the effect of a covariate that takes into account whether respondents that are
concerned with local effects of climate change have a higher WTP compared to respondents
not particularly concerned with local effects. We decided to vary the level of information
provided to respondents about the effects of climate change and mitigation programmes to
assess whether respondents perceived the local benefits of climate change only when they
were highlighted about those effects. We therefore split the sample of respondents into two
subsamples that faced different levels of information in the description of the programs: one
half of the respondents was presented with the climate policies in terms of global b enefits
and the other half was presented with information both on global benefits and on the local
effects in the Basque Country.
The scenarios presented to the respondents read as follows (in italics the additional infor-
mation about the local benefits and damages shown only to half of the respondents):
“In recent decades, people have been putting large amounts of “greenhouse gases” such
as carbon dioxide into the air, mostly by burning gasoline, coal and natural gas. Scientists say
that as a result the Earth has been getting warmer and that it will continue to get warmer in the
future. They have warned that average world temperatures might increase by approximately
3
◦
C by 2050. If the Earth keeps getting warmer and our climate changes, this could hurt
people and nature: scientists say that there are likely to be more droughts, flooding of coastal
areas where people live, more severe storms, plants and animals will become extinct and
many diseases will spread. Global warming could also be beneficial for some areas of the
Earth. Scientists also expect the following consequences for the Basque Country by the end
of the twentyfirst century: average summer temperatures should increase as much as 5.5
◦
C,
more irregular rainfall will increase the risks of flooding near the coast; the sea level would
rise of about 40 cm causing a loss of beach areas, infrastructures and constructions, and
an increase in sea storms. Expected effects on human health in the Basque Country include
an increase in rates of sickness and death due to heat waves, increased episodes of acute
respiratory problems such as pulmonary infections and serious asthma attacks especially
among children and the elderly. Possible positive effects of global warming in the Basque
Country could include a lower use of heating in winter.”
Similarly, when respondents were presented the hypothetical programs, two different lev-
els of information were offered to the two groups of respondents. For example, for the BPCCC
scenario with a tax of t e, respondents were told that:
“The previous two policies would target only 4.5% of all the greenhouse gases emissions
of the Basque Country. A recent study by the University of the Basque Country has shown
123
126 A. Longo et al.
that the total cost of reducing the greenhouse gases emissions of the Basque Country by 16%
by 2012 compared to the current emissions trend to comply with the Kyoto Protocol targets
is equal to 24 Million e or t e for each household in the Basque Country. This cost of t e
per household includes the costs of the two measures for renewable electricity and energy
efficiency previously described and other additional measures such as improving the control
of emissions from industry and agriculture, reducing the emissions from private and public
transport and restructuring several public buildings to make them more energy efficient.
The Basque Country government is therefore considering, instead of the previous two
policies on renewable electricity and energy efficiency, to impose a climate change tax of
t e that each household in the Basque Country should pay for four years from 2009 till
2012. This tax would decrease emissions by 16% by 2012 compared to the current emissions
trend, improve local air quality and decrease the number of respiratory problems such as
pulmonary infections and serious asthma attacks.”
Finally, the questionnaire also included some questions on individual attitudes towards
global warming. In one of these questions, respondents were asked about the relative impor-
tance that they gave to the effects of climate change on the Basque Country. Answers to this
question are used to assess whether respondents who consider the effects of climate change
to be important for the Basque Country have a higher WTP compared to other respondents
who are mostly concerned for the global effects of climate change.
4 Economic and Econometric Models
As described in section three, each respondent faced three dichotomous choice WTP ques-
tions for implementing three programmes to cut GHG emissions in the Basque Country. For
each programme, respondents were asked to state whether they were willing to pay a pro-
posed tax to fund the measure. Answers to these questions are analyzed within the random
utility model framework (Hanemann 1984) that assumes that respondents would choose to
pay the proposed tax if the utility they would receive from paying the tax would be higher
or equal to the proposed tax. For good m, respondent i would be willing to pay the tax t
m
if
her utility is higher with good m, than without it:
U
i
= U
im
(m, Z, −t
m
) ≥ U
i
(Z), (1)
where Z represents the other goods and services consumed by respondent i. As there may be
elements in (1) not observable to the researcher, it is necessary to include an error term. For
the econometric analysis, sequential equations can each be estimated consistently by means
of individual single equation probit models. However, this procedure would be inefficient
because it ignores the possible correlation between the disturbances (Poe et al. 1997). In
the presence of sequential CV questions, joint estimation of CV responses provides more
precise WTP estimates (Park and Loomis 1996). In this context, a multivariate probit model
is a generalisation of the probit model that allows for the joint estimation of several correlated
binary outcomes. The general specification for a multivariate probit model is:
Y
∗
im
= α
m
T
im
+ β
m
X
im
+ ε
im
m = 1, 2,...,M,
Y
im
= 1ifY
∗
im
> 0 and 0 otherwise, (2)
where Y
∗
im
is an unobserved variable that represents the latent utility or propensity of accept-
ing the proposed bid by individual i in question m (in our case, m = 1, 2, 3) and is equal to
1 if individual i answers ‘yes’ to the WTP question, and 0 otherwise; T
im
is a vector of tax
123
Willingness to Pay 127
amounts presented for programme m to respondent i in the CV question m; X
im
is a matrix
of observed variables considered relevant for explaining the choice of individual i for voting
in favour or against the proposed tax in question m; α
m
and β
m
are vectors of coefficients
to be estimated for each question m;andε
im
is an error term assumed to be normally (0,1)
distributed. The variance-covariance matrix of the error term is thus (for m = 3):
=
⎡
⎣
1 ρ
12
ρ
13
ρ
12
1 ρ
23
ρ
13
ρ
23
1
⎤
⎦
. (3)
The probability of observing a ‘yes’ response conditioned on α, β, , T and X can be written
as follows (Chib and Greenberg 1998):
Pr(Y
i
= y
i
, i = 1, 2, 3|α, β, ) =
A
1
A
2
A
3
φ(Z
1
, Z
2
, Z
3
,ρ
12,
ρ
13,
ρ
23
)dz
3
dz
2
dz
1
, (4)
where φ is the density function of a multivariate normal distribution with mean vector 0 and
the variance-covariance matrix , Z
1
, Z
2
, and Z
3
are matrixes blocked by vector T
m
and
matrix X
m
and A
i
is the interval (−∞,β
m
X
im
) if y
i
= 1and(β
m
X
im
, ∞) if y
i
= 0.
The mean WTP for programme m with respect to preference uncertainty (ε) can be
obtained from the following formula:
E
ε
(WTP
m
|α, β, X
i
) =−
β
m
Z
im
α
m
. (5)
Our Hypothesis I (see Table 2) is that ancillary benefits in climate change mitigation poli-
cies do matter. For this purpose, we add a dummy variable (ANCIL_INFO) taking on the
value of 1 when respondents face the questionnaire version with the description of the local
effects of climate change and climate change mitigation measures, in addition to the informa-
tion of the global effects, and 0 when respondents face the questionnaire without the ancillary
benefits information. A positive and significant sign on the ANCIL_INFO coefficient would
suggest that respondents that were highlighted with the local effects of climate change and
climate change mitigation measures have a higher WTP compared to those respondents that
were not given such information. A not significant coefficient would suggest that respon-
dents’ WTP was not affected by the amount of information received. Therefore, to properly
assess the WTP for ancillary benefits, we add a dummy variable (ANCIL_LOCAL) equal to
1 to account for respondents that thought that climate change was an extremely important or
very important issue for the Basque Country, and equal to 0 for those that did not consider
Table 2 Hypotheses tested with the model
Hypothesis Description
I Ancillary benefits of climate change mitigation measures matter
II WTP measures are internally valid
III WTP measures pass the scope test
IV WTP is higher when accounting for energy security
V WTP is higher for those having Basque cultural identity
VI Members of environmental NGOs, education level and left wing political views positively
affect the WTP for climate mitigation policies
123
128 A. Longo et al.
important the effects of climate change for the Basque Country. A positive and significant
coefficient for ANCIL_LOCAL would suggest that respondents that are concerned also for
the effects of climate change in the Basque Country have a higher WTP due to the ancillary
benefits of climate change mitigation measures.
As a standard procedure to check the validity of the survey instrument, the second hypoth-
esis (Hypothesis II) of our model aims to analyse the internal validity of our responses. To
do so, we include a dummy variable (HIGHI) taking on the value of 1 if a respondent has a
high income.
3
The third hypothesis (Hypothesis III) of our model is that the CV study passes the ‘scope’
test. Theoretically, when valuing two nested goods, economic theory suggests that the larger
good should be valued the same or higher than the smaller good, although there is empirical
evidence showing scope insensitivity.
4
Given that in our survey, the BPCCC programme
nests the other two programmes for promoting renewable energy and for improving energy
efficient measures, we would expect that the WTP(energy efficiency) ≤ WTP(renewable
electricity) ≤ WTP(BPCCC).
Hypothesis IV aims to examine the importance of energy security concerns on the WTP
for climate change mitigation policies. Policies aimed at guaranteeing energy security are of
interest to the European Commission and national governments and they have been found
to be supported by energy consumers concerned with climate change (Longo et al. 2008).
To test for the importance of energy disruptions, we create the dummy variable BLACK
equal to 1 when an individual is extremely or very concerned with the frequency of energy
blackouts, and 0 otherwise.
Hypothesis V aims to test whether cultural identity is a significant explanatory variable
of the WTP to protect natural resources, as it has been highlighted by Hoyos et al. (2009). A
dummy variable (IDENT) controls for respondents’ self-reported Basque cultural identity.
This variable allows us to investigate whether respondents reporting a high level of Basque
cultural identity, as they may be more closely interested in the local effects of policies related
to the Basque Country, may also have a higher WTP. A positive and significant sign for this
variable would provide additional indirect evidence of the importance o f ancillary benefits
of climate change mitigation measures.
Finally, we wish to test (Hypothesis VI) whether some respondents’ characteristics found
to be significant in the literature (i.e. level of education, membership in an environmental
organisation and left wing political views) influence the WTP for climate change mitigation
by adding respective covariates to the model. Previous research on the effect of membership
in an environmental organisation on environmental WTP has found a positive influence for
voting in favour of climate change mitigation programs (e.g. Longo et al. 2008). Research on
the effects of education on WTP has found mixed results. On the one hand, Blomquist and
Whitehead (1998), Witzke and Urfei (2001)andLi et al. (2004) have reported a positive rela-
tion between level of education and environmental WTP. On the other hand, Danielson et al.
(1995), Krupnick et al. (2002), Bergmann et al. (2006)andLongo et al. (2008) have found
the opposite effect. Finally, studies by Berrens et al. (2004), Veisten et al. (2004)andPopp
(2001) have found that the level of education has no significant effect on WTP. In respect
to political views, Bannon et al. (2007) find evidence that policies aiming at reducing GHG
emissions in the United States were more favoured by Democrats. We deem it interesting to
investigate how political views affect WTP for climate change mitigation strategies in the
3
Income level was coded in 5 levels (1 being low income and 5 high income level) and it was converted into
a dummy variable taking on the value of 1 if the respondent was in levels 4 and 5, and 0 otherwise.
4
Scope insensitivity has been raised as a major argument against the reliability of CV (Arrow et al. 1993).
123
Willingness to Pay 129
Basque Country, as, to our knowledge, no previous empirical study using the CV method for
valuing climate change mitigation programs has looked into this aspect.
5 Results and Discussion
5.1 Descriptive Statistics and Initial Questions
Before analysing the answers to the WTP questions, we check how well our sample repre-
sents the population of the Basque Country. We do so by comparing our data with the official
statistics of the Basque Country and other large household surveys recently conducted by the
government. Our final sample of 1,000 completed questionnaires well matches the official
statistics in terms of age, gender, education, occupation, family size and income compared
to the official statistics for the Basque Country (EUSTAT 2008; INE 2008).
5
The first set of questions in the survey instrument investigates how important environ-
mental problems are perceived to be and what the respondents’ knowledge of and attitudes
toward climate change are. The purpose of these questions is to let respondents feel com-
fortable with answering general questions—‘warm up’ questions—and to assess the order
of importance of climate change mitigation measures compared to other public priorities.
Table 3 shows that when asked to choose the most important goal for the Basque Country
today, respondents consider economic growth and employment the most important goals,
followed by reducing political and social conflicts, decreasing poverty and then protecting
the environment. In 10 years time, respondents consider that economic growth and employ-
ment will still be important, but less than they are today, whilst environmental problems will
be the second most important problem, followed by reducing political and social conflicts.
Comparing these results with the answers obtained by the Ecobarometro 2008 survey,
6
we
observe that after the recent ‘credit crunch,’ respondents are much more concerned with both
the economic growth and the environment than the respondents that answered the same ques-
tion one year in advance for the Ecobarometro survey. We find that 42% of respondents claim
that economic growth and employment are important, whilst only 32% had the same view in
the Ecobarometro 2008 survey. The respondents of our survey are also less concerned with
decreasing poverty than the respondents of the Ecobarometro survey were: only 14% of our
respondents suggest that the most important goal is to decrease poverty, compared to 21%
of the Ecobarometro survey. We read these results as showing that the recent ‘credit crunch’
has probably lead people to be less altruistic and more worried about the economy. What is
most interesting to us is that our results are consistent with the Ecobarometro findings when
considering the protection of the environment: we (Ecobarometro) find that 10% (12%) and
26% (26%) of respondents consider the protection of the environment as the most impor-
tant goal today and in 10 years respectively. These results highlight how, despite the recent
5
Descriptive statistics for our survey compare to official statistics reported in parenthesis as follows: per-
centage of males is 45% (48.84%); average age is 40.15 (41.29) years; the average number of adults in the
household is 2.60 (2.51); the average number of children living at home is 0.51 (0.41); the unemployment rate
is 4.90% (5.60%); the percentage of student is 10.51% (12.27%); the percentage of respondents with basic
education is 31.12% (33.14%); personal monthly income is e1,119.26 (e1,137.25).
6
The Ecobarometro Social (Social Ecobarometer) is a periodical publication aiming at monitoring the evo-
lution of the Basque public opinion about the environment. For the Ecobarometro Social 2008, conducted
in November 2007, 2000 home interviews at a representative sample of the population were carried out.
Ecobarometro Social is part of the European Eurobarometer surveys, co-ordinated by the European Commis-
sion (Euskobarometro 2008).
123
130 A. Longo et al.
Table 3 Most important goals for the Basque Country today, and in 10 years
Our survey Ecobarometro 2008
Today
(N = 1,000) %
In 10 years
(N = 1,000) %
Today
(N = 2,000) %
In 10 years
(N = 2,000) %
Economic growth and
employment
42.20 31.10 32 23
Decrease poverty 14.10 12.40 21 14
Protect the environment 10.00 26.30 12 26
Reduce political and social
conflicts
24.80 15.60 20 13
Improve public health 4.90 7.60 8 12
Give more participation to
people on important
decisions (government, city
council)
2.00 3.00 4 4
Other 1.40 1.40 2 2
Don’t know 0.60 2.60 2 7
financial crisis, the protection of the environment remains an important goal and it will be
more and more important in the coming years.
Answers to other initial questions of the survey show that a majority of Basques believes
that climate change is occurring (93%) and they are willing to change their habits to be more
environmentally friendly (53%). Results show that the respondents are generally aware of
the existence of human induced climate change, with only 3.5% of the respondents stating
that climate change is a natural process of the Earth. These results are in consonance with
similar studies in Spain (Fundación 2008) and the United States (Bannon et al. 2007).
After acknowledging that the characteristics of our sample are comparable with those
of the population of the Basque Country, we focus on the results from the CV questions.
A complete description of the variables used in the econometric models estimation along
with their descriptive statistics is provided in Table 4. Potential problems or correlation or
multicollinearity were examined but none of them was found.
5.2 Model Specification and Estimation Results
In this section, the answers to the WTP questions for the three programmes are analysed.
The data were first analysed using binary probit models, and then, to account for the panel
structure of our dataset and for the interdependence in responses across alternative valuation
scenarios, a multivariate (trivariate in this case) probit model was used. Maximum likelihood
parameters and correlations of the error terms were estimated using LIMDEP econometric
software (Greene 2007). As the multivariate probit model outperforms single equation probit
models, suggesting that our data do contain respondent specific effects, we only report the
results from the multivariate probit model.
7
After having checked that deleting 214 protest
respondents—respondents who did not engage in the CV exercise—does not lead to item
7
Coefficient estimates from single equation probit models provides almost identical output to the estimates
from the multivariate probit model. However, the multivariate probit model offers a better fit of the data. For
example, for model 1 in Table 5 the gain in the log likelihood function from the multivariate probit model is
of 236 compared to the sum of the log likelihood functions of the single equation probit models.
123
Willingness to Pay 131
Table 4 Socioeconomic variables and summary statistics
Variable Definition Mean SD Min Max Observations
AGE Age of respondents 39.66 16.09 18 88 738
MALE Percentage of male respondents 0.45 0.50 0 1 738
BIDRE Tax amount for the renewable electricity
program
68.25 59.26 10 180 738
BIDEE Tax amount for the energy efficiency
program
53.41 51.16 5 150 738
BIDCC Tax amount for the BPCCC program 134.93 116.80 20 350 738
NGO Dummy variable equal to 1 if respondent
is member of an environmental
organisation; and 0 otherwise
0.04 0.19 0 1 738
IDENT Dummy variable equal to 1 if respondent
has Basque cultural identity; and 0
otherwise
0.18 0.38 0 1 738
LEFT Dummy variable equal to 1 if respondent
tends to favour left political views; and
0 otherwise
0.29 0.45 0 1 738
LSTUD Dummy variable equal to 1 if the level of
education of the individual is low; and
0 otherwise
a
0.16 0.36 0 1 738
HIGHI Dummy variable equal to 1 if respondent
has high income level; and 0 otherwise
0.33 0.47 0 1 738
ANCIL_INFO Dummy variable equal to 1 if respondent
faces the questionnaire version with the
description of the local effects of
climate change and climate change
mitigation measures
0.51 0.50 0 1 738
ANCIL_LOCAL Dummy variable equal to 1 to if
respondent thinks that climate change
is an extremely or very important issue
for the Basque Country; and 0
otherwise
0.68 0.47 0 1 738
BLACK Dummy variable equal to 1 if respondent
is extremely or very concerned with the
frequency of energy blackouts; and 0
otherwise
0.90 0.30 0 1 738
a
Educational level was codified coded in 5 levels (1 being low and 5 high educational level) and it was
converted into a dummy variable taking on the value 1 if the respondent was in levels 1 and 2, and 0 otherwise
non response bias, as we found that protest respondents were not systematically different
from non-protest respondents, we ran the econometric models on the sample cleaned from
protest respondents (Freeman 2003).
8
The results from a parsimonious specification of the multivariate probit model that explains
the answers to the payment questions with the bid vector and a constant term for each equa-
tion are reported in Model 1 (see Table 5).
9
This model shows that the three coefficients
8
We ran a two sample t-test of equality of mean values for income, gender, marital status, employment, level
of education, cultural identity and found that we could not reject at the 5% significance level that the mean
values are equal across protest and non protest respondents.
9
We tested whether our data suffer from ordering effects by entering a dummy variable equal to 1 if respon-
dents received the questionnaire with first the CV question on the renewable electricity program followed by
the CV question on the program on energy efficiency and 0 if respondents received the version of the ques-
tionnaire where first they answered the CV question on energy efficiency followed by the one on renewable
123
132 A. Longo et al.
Table 5 Multivariate probit models results
Model 1 Model 2 Model 3
Loglikelihood function −1093.045 −1081.614 −1061.304
Coeff t-stat Coeff t-stat Coeff t-stat
Renewable electricity equation
Constant renewable energy 0.9502 12.21 0.6509 5.71 0.0958 0.37
BIDRE −0.0054 −6.77 −0.0054 −6.64 −0.0055 −6.45
ANCIL_INFO 0.0518 0.51
ANCIL_LOCAL 0.4075 3.87 0.3625 3.30
NGO 0.3953 1.23
HIGHI 0.1031 0.87
IDENT 0.1490 1.06
LEFT 0.3640 3.06
LSTUD −0.1613 −1.12
BLACK 0.5289 3.17
Energy efficiency equation
Constant energy efficiency 0.9233 12.38 0.6295 5.55 0.0939 0.
47
BIDEE −0.0071 −7.46 −0.0071 −7.44 −0.0071 −6.93
ANCIL_INFO 0.0518 0.41
ANCIL_LOCAL 0.4075 3.87 0.3536 3.14
NGO 0.4446 1.25
HIGHI 0.0309 0.26
IDENT 0.2425 1.70
LEFT 0.2754 2.45
LSTUD −0.1613 −1.12
BLACK 0.5289 3.17
BPCCC equation
Constant BPCCC 0.5317 7.30 0.4521 3.93 0.2607 1.37
BIDCC −0.0019 −4.73 −0.
0019 −4.67 −0.0019 −4.68
ANCIL_INFO −0.1829 −1.91
ANCIL_LOCAL 0.2593 2.54 0.2094 1.97
NGO 0.7121 2.49
HIGHI 0.1821 1.70
IDENT −0.1141 −0.89
LEFT 0.2066 1.86
LSTUD −0.1061 −0.77
BLACK 0.0323 0.20
Corr(renewable energy,
energy efficiency)
0.9101 49.47 0.9067 47.44 0.9032 43.90
Corr(renewable energy,
BPCCC)
0.6142 13.46 0.6027 12.78 0.5979 12.08
Corr(energy efficiency,
BPCCC)
0.4974 9.53 0.4859 9.11 0.4826 8.80
123
Willingness to Pay 133
Table 5 continued
Model 1 Model 2 Model 3
Loglikelihood function −1093.045 −1081.614 −1061.304
Coeff t-stat Coeff t-stat Coeff t-stat
Mean
WTP
Standard
error
Mean
WTP
Standard
error
Mean
WTP
Standard
error
WTP for renewable energy
(e)
173.27 17.33 176.03 17.88 176.24 18.33
WTP for energy efficiency
(e)
129.92 11.85 131.46 12.38 132.01 9.61
WTP for BPCCC (e) 279.08 38.94 280.79 39.60 281.61 40.12
Dependent variables are the probability of choosing “yes” to the WTP questions. From the top of table to the
bottom, we first report the output for the probability of choosing “yes” to the CV question for the renewable
electricity programme; followed by the probability of choosing “yes” to the CV question for the energy effi-
ciency programme; followed by the probability of choosing “yes” to the CV question for implementing the
BPCCC. Standard errors for WTP are calculated with the Delta method. Number of observations is 738 after
deleting protesters and missing observations
for the constant terms are positive and significant, suggesting that the respondents support
the proposed climate mitigation policies, and the three coefficients for the tax amounts are
negative and significantly different from zero, suggesting that utility decreases if the cost of
the programme increases (a different but equivalent interpretation is that respondents are less
likely to choose a programme the more expensive it is, other things being constant). Esti-
mates of mean WTP are also reported in Table 5, showing that the respondents have a mean
WTP of e173.27, e129.92 and e279.08 for implementing the programmes for supporting
renewable electricity, energy efficient measures in the house and the BPCCC respectively.
Finally, the output from Model 1 reports a strong positive correlation between the individual
probit equations. Positive correlation among the three equations suggests that when a respon-
dent supports one programme, she is also more likely to support the other programmes. We
specifically find a very strong and positive correlation between the equations for the WTP
for renewable electricity and for the WTP for energy efficiency (the correlation coefficient is
equal to 0.91), suggesting that most respondents (not) supporting one programme would also
(not) support the other one. The correlations between the WTP equations for the BPCCC and
the other two programmes are still positive and significant but smaller than the correlation
between the two partial programmes. The correlation coefficients are equal to 0.61 between
the equations for the BPCCC and the renewable electricity and 0.49 for the equations for
the BPCCC and energy efficient measures. The magnitude of the coefficients and the fact
that they are statistically different from zero supports the choice of the multivariate probit
model as a correct model to analyse the data. The results conform with the literature on CV, as
Poe et al. (1997) suggest that nested goods are likely to result in a positive correlation between
the error terms.
A second specification, Model 2, aims at investigating the WTP for ancillary benefits in
climate change mitigation policies. It explains the answers to the payment questions by add-
ing to the previous specification two dummy variables (ANCIL_INFO and ANCIL_LOCAL)
Footnote 9 continued
electricity. As we found that such a dummy variable was not significantly different from zero at the 10%
significance level, we did not include it in our models.
123
134 A. Longo et al.
in each equation. A likelihood ratio test suggests that including these variables significantly
improves the fit of the model (LR = 22.86 >χ
2
6
= 12.59). The non-significant coefficient
of the variable ANCIL_INFO suggests that adding information on local benefits of climate
change policies has no influence on respondent’s WTP answers.
10
On the contrary, the coef-
ficient for the ANCIL_LOCAL variable is significant and positive, suggesting that utility
increases if ancillary benefits are considered: respondents that are concerned for ancillary
effects of climate change are more likely to support the programmes for reducing GHG
emissions and have a higher WTP for the programmes compared to those that are not con-
cerned with local effects of climate change and climate change mitigation measures. Fur-
thermore, these results for the ANCIL_INFO and ANCIL_LOCAL coefficients suggest that
the consideration of ancillary benefits is not ‘informationally’-induced but an intrinsic part
of respondents’ WTP. In fact, the results are robust in two ways: the questionnaire is robust
because providing more information with emphasis on the local effects of climate change
and climate change mitigation programmes does not affect WTP values; ancillary benefits
are not affected by what the respondents are told in the description of the programmes, sug-
gesting that those respondents that care for ancillary benefits do so irrespectively of what
they are told (i.e. there is no information bias). T herefore, Model 2 provides us with reasons
not to reject Hypothesis I: ancillary benefits positively affect the WTP for climate change
mitigation policies and therefore they do matter.
Model 3 incorporates to the specification of Model 1 the influence of socioeconomic
characteristics of the respondents: membership in environmental organisations (NGO), high
income level (HIGHI), cultural identity (IDENT), left wing political views (LEFT), low edu-
cational level (LSTUD) and the consideration of energy disruptions (BLACK). Note that the
fit of this model is again significantly higher than the fit of Model 1, as shown by the cor-
responding likelihood ratio test (LR = 63.48 >χ
2
21
=32.67). All the coefficients in the three
equations have the expected signs, but their significance level varies across equations. Here
we only discuss the coefficients that are significant. Firstly, a respondent is more likely to sup-
port the programme on renewable electricity if she considers ancillary benefits, is politically
left-winged oriented or concerned about energy security issues. Secondly, a respondent is
more likely to support the programme on energy efficiency if she considers ancillary benefits,
her cultural identity is Basque, is left-winged oriented or concerned about energy security
issues. And thirdly, a respondent is more likely to support the whole BPCCC if she considers
ancillary benefits, is member of an environmental organisation, has a high income level and
is left-winged oriented. The level of correlation between the three probit models is similar to
those observed in Models 1 and 2.
5.3 Willingness to Pay
In this section, the heterogeneity of respondents’ WTP values is further examined using
model 3 and Eq. 5 to predict individual WTP using respondents’ characteristics. We then
calculate the mean WTP values for two groups: a base group—respondents who do not have
the characteristic of interest—and a c ontrol group—respondents who do have such char-
acteristic. In this way we can account for differences in mean WTP for the characteristic
of interest holding constant all other characteristics. For example, as reported in Table 6,
10
As pointed out by an anonymous reviewer, the lack of statistically significance of the ANCIL_INFO coef-
ficient estimate may be due to the fact that the information o n local benefits and damages described in the
hypothetical scenario is quite general. Such information, however, is the exact information used in the offi-
cial documents by the Basque Government (BG 2008). In addition, focus groups participants were able to
understand the local impacts as were described in the questionnaire.
123
Willingness to Pay 135
Table 6 WTP for climate change mitigation
Implementation of
the BPCCC
Promotion of
renewable energy
Promotion of
energy efficiency
Base
group
Control
group
%
change
Base
group
Control
group
%
change
Base
group
Control
group
%
change
Ancillary benefits 186.35 326.71 75.32 124.74 200.30 60.57 94.60 149.45 57.99
Environmental
NGO 266.08 686.01 157.82
High income 243.47 361.99 48.68
Left wing politics 241.35 380.94 57.84 154.83 228.32 47.47 119.36 162.70 36.32
Energy concerns 86.82 186.36 114.65 62.72 139.82 122.94
Cultural identity 128.63 147.68 14.82
Base group: mean WTP values for respondents not having the characteristic of interest
Control group: mean WTP values for respondents having the characteristic of interest
the mean WTP for implementing the BPCCC is equal to e186.35 for respondents that did
not consider ancillary benefits and e326.71 for respondents that did consider the ancillary
benefits, holding constant their political views, cultural identity, energy concerns, income
level and being member of an environmental NGO. The mean WTP for promoting renew-
able electricity is estimated at e124.74, increasing to e200.30 when respondents care about
the ancillary benefits of climate change programs. Similarly, the mean WTP for the energy
efficiency programme is estimated at e94.60 and it increases to e149.45 when respondents
consider the ancillary benefits of climate change programs. In other words, WTP for climate
change mitigation measures is 75, 60 and 58% higher for the three programmes when ancil-
lary benefits are considered. These results are in line with the lower bound estimates of the
ancillary benefits assessed through macroeconomic models (Pearce 2000).
Table 6 also shows that (1) membership of an environmental organisation increases the
estimated WTP for the implementation of the BPCCC by 158%, (2) respondents with high
income level have a WTP for implementing the BPCCC 49% higher, (3) left wing political
views increases estimated WTPs by 48% in the case of the policy for promoting renewable
electricity, 36% in the case of the policy for promoting energy efficiency and 58% in the case
of the policy for implementing the BPCCC, (4) Basque cultural identity of the respondents
increases their estimated WTP by 15% in the policy for promoting energy efficiency, and (5)
energy security concerned people’s WTP is 115% higher if a policy for promoting renewable
electricity is implemented and 123% higher if a policy for promoting energy efficiency is
implemented.
The relationship between income and WTP is usually analysed to test for internal validity
of the survey (Bateman et al. 2002). In this case, we find that respondents with higher income
are more willing to support the BPCCC. We can therefore conclude that our WTP estimates
are internally valid and that Hypothesis II cannot be rejected. These results also conform
with the literature on CV and provide evidence in support of Hypothesis III: we find that
WTP increases for programmes offering a higher reduction in GHG emissions and that the
WTP is greater for the options that result in greater cuts in GHG. Scope effects were tested
by measuring the difference in mean WTP using the test for non-independent distributions
proposed by Poe et al. (1997). The p-value for the null hypothesis that the E[WTP(BPCCC)]
≤ E[WTP(renewable energy)] was < 0.001, and the p-value for the null hypothesis that the
123
136 A. Longo et al.
0
0.01
0.02
0.03
0.04
0 100 200 300 400 500 600
E[WTP]
Prob
ENERGY_EFFICIENCY
RENEWABLE_ELECTRICITY
BPCCC
Fig. 2 Simulated WTP for climate change mitigation
0.00
0.00
0.01
0.01
0.02
0.02
0 50 100 150 200 250 300 350 400 450
WTP(BPCCC)-WTP(RENEWABLE_ENERGY)
Frequency
0.00
0.01
0.02
0.03
0.04
0.05
-20 0 20 40 60 80 100 120 140 160
WTP(RENEWABLE_ENERGY)-WTP(ENERGY_EFFICIENCY)
Frequency
Fig. 3 CDF of the differences of the mean WTP distributions
E[WTP(renewable energy)] ≤ E[WTP(energy efficiency)] was 0.002. We cannot therefore
reject the hypotheses that our WTP estimates pass the scope test (see also Figs. 1, 2, 3).
Hence, the result that WTP(BPCCC) > WTP(renewable energy) > WTP(energy efficiency)
confirms that this research passes the scope test: respondents are willing to pay more money
for a programme that offers larger cuts in greenhouse gases emissions. Similar results were
found in Layton and Brown (2000).
Another interesting result is that the WTP for the BPCCC is smaller than the sum of the
two other programmes for reducing GHG emissions from promoting renewable electricity
and from encouraging more energy efficient systems in the house. This outcome had previ-
ously been observed also by Carson and Hanemann (2005, p. 910) that point out that “when
the goods being valued are normal goods and (Hicksian) substitutes for each other, the value
of a particular public good should be progressively smaller the later in a WTP sequence it is
valued.”
123
Willingness to Pay 137
The positive and significant signs of the coefficient BLACK for the renewable electricity
and energy efficiency measures equations provide support for Hypothesis IV, suggesting that
respondents concerned with black outs are willing to pay more for the two nested public
programmes that reduce the emissions of GHG. Table 4 also provides support in favour of
Hypothesis V: similar to the results reported in Hoyos et al. (2009), we find that Basque
cultural identity positively affects the WTP for the environmental programmes. However,
only for the equation for energy efficient measures the coefficient of IDENT is significant.
Therefore, we conclude that having Basque cultural identity increases the mean WTP for
climate change mitigation by promoting energy efficiency at home by 15%.
Finally, with regards to Hypothesis VI we find mixed results. Firstly, left wing political
views positively affect WTP for the three programmes, reflecting the political situation in
Spain where the left wing Socialist Party has been more actively in favour for actions against
climate change as compared with the right wing Popular Party. Similar results were reported
by Bannon et al. (2007) in the US, where Democrats were more likely than Republicans to
favour climate change mitigation policies. Secondly, being members of an environmental
NGO positively affect WTP for the BPCCC (but not for the two nested programmes, as the
coefficient for NGO is not significant for the renewable energy and the energy efficient mea-
sures equations). These results are similar to those obtained by Longo et al. (2008). However,
we find no e ffect of the level of education on the WTP for climate change mitigation policies,
as the coefficients for this variable are not significant in the three equations.
11
These results
are similar to those obtained by Berrens et al. (2004), Veisten et al. (2004)andPopp (2001).
6 Conclusions
This study has examined the preferences and attitudes towards climate change of the popu-
lation of the Basque Country by interviewing a representative sample of 1,000 households
using a Contingent Valuation questionnaire. Results to the attitudinal questions show that the
respondents are generally aware of the existence of human induced climate change and are
concerned with its effects at global and local scale, in addition to concern for its consequences
on future generations. The CV was used to elicit the WTP for global and ancillary benefits
of climate mitigation policies under the recently approved BPCCC. Basque households were
asked for their WTP for (1) promoting renewable energy aiming at reducing GHG emissions
by 4%, (2) improving energy efficient measures at home, aiming at reducing GHG emissions
by 0.5% and (3) and for implementing the BPCCC, aiming at achieving the emission targets
agreed in the Kyoto Protocol by reducing GHG emissions by 16%. Following the litera-
ture on multiple CV questions, single bounded dichotomous choice CV questions were used
and models were estimated accounting for interdependence in responses across alternative
valuation scenarios. When we looked at the WTP for implementing the BPCCC, we found
that respondents favour the implementation of the plan and are willing to pay an increase
in taxes to fund the programme. The results of the CV questions reveal that respondents are
on average willing to pay an additional annual tax of 281.61 e (s.e. 40.12) for a four year
period to fund projects aimed at attaining the reductions in GHG emissions required under
the BPCCC (16% reduction in GHG emissions). Respondents were on average willing to pay
an additional annual tax of 176.24 e (s.e. 18.33) for implementing a programme for support-
11
In a specification not reported here, we checked whether a model with dummy variables for each level of
education would improve the fit of the model, but it did not.
123
138 A. Longo et al.
ing renewable energy (4% reduction in GHG) and 132.01 e (s.e. 9.61) for implementing a
programme for supporting energy efficient measures at home (0.5% reduction in GHG).
These results are policy relevant. From a policy perspective, they show that Basque citi-
zens are willing to support the implementation of the BPCCC through the introduction of a
new tax. The social benefits of implementing the Plan, estimated at 850.66 million e,
12
is
much greater than its costs of implementation, estimated at 79.50 million e (BG 2008), so the
Plan would have net social benefits. Furthermore, ancillary benefits represent 34% (288.81
million e) of the social benefits derived from implanting the BPCCC. In other words, one
third of the perceived benefits can be attributed to the ancillary benefits of climate change
mitigation policies.
Our study further aimed at disentangling ancillary benefits from global benefits of climate
change mitigation. The consideration of ancillary benefits has been found to significantly
increase the probability of voting in favour of climate mitigation policies. On average, esti-
mated WTP for climate change mitigation policies was increased by 58–75% when ancillary
benefits were considered. If these results are converted into a figure per ton of CO
2
abated,
we find that individuals are willing to pay 57 e (84.39$) /ton CO
2
abated in the case of the
policy for promoting renewable electricity, 332 e (491.54$) /ton CO
2
abated in the case of
the policy for promoting energy efficiency and 26 e (39.19$) /ton CO
2
abated in the case of
the implementation of the BPCCC.
13,14
We could say that the differences in WTP between
alternatives reflect also the perception of ancillary benefits. These results conform with the
estimates from the macroeconomic models reviewed by the OECD (2000), and highlight that
the WTP for CO
2
abatement depends on the abatement scheme used.
Other major results of interest from our study are the following: (1) the research is inter-
nally valid and passes the scope test; (2) black-outs avoidance concern increases estimated
WTP for promoting renewable electricity and energy efficiency; (3) Basque cultural iden-
tity is positively related to environmental WTP; and (4) membership of an environmental
NGO and left wing political views positively affect the WTP. So, in line with Layton and
Brown (2000), the results show, on the one hand, that people care about long-term effects
of climate change that in many cases they may not even suffer and, on the other hand,
that preferences regarding climate change mitigation are very heterogeneous. Finally, our
results are policy relevant as they provide useful information to the Basque government on
how much the general population is willing to pay for abating GHG emissions, how much
people are concerned about local and personal consequences of climate change compared to
global consequences, and which levels of different policy instruments are acceptable by the
public for implementing a climate change mitigation strategy. Future research is needed to
address different ways of capturing ancillary benefits of climate mitigation policies using CV
studies.
Acknowledgements The authors acknowledge the financial support from the Department of Environment
of the Basque Government and from the Department of Education of the Basque Government through grant
12
Social benefits are calculated considering the number of households in the Basque Country and a social
discount rate of 5%.
13
The WTP for ton of CO
2
abated was calculated using the estimates from Table 6, by subtracting the WTP of
respondents who did not consider ancillary benefits from the WTP of those that considered ancillary benefits,
multiplying it by the number of households in the Basque Country (811,311), and then dividing it for the tons
of CO
2
abated from implementing the specified policy (4,300,000 for the BPCCC, 1,075,000 for the renewable
energy program, 134,000 for the energy efficiency program). For example, the WTP / ton CO
2
abated from
renewable electricity is equal to: (200.302 − 124.741) * 811,311/1,075,000 = 57.03e.
14
The Euro was worth 1.48 US dollars at the time of the survey.
123
Willingness to Pay 139
IT-334-07 (UPV/EHU Econometrics Research Group). We would also like to thank Gregory Poe and referees
of this journal for their comments. Full responsibility of errors remains with the authors.
References
Alfsen KH, Brendemoen A, Glomsrød S (1992) Benefits of climate policies: some tentative calculations.
Discussion paper no. 69, Central Bureau of Statistics, Oslo
Arrow K, Solow R, Portney P, Leamer E, Radner R, Shuman H (1993) Report of NOAA panel on contingent
valuation. Fed Regist 58:4601–4614
Ayres RU, Walter J (1991) The greenhouse effect: damages, costs and abatement. Environ Resour Econ 1:
237–270
Bannon B, DeBell M, Krosnick J, Kopp R, Aldhous P (2007) Americans’ evaluation of policies to reduce
greenhouse gas emissions. National Press Club, Washington, DC
Barker T, Johnstone N, O’Shea T (1993) The CEC carbon/energy tax and secondary transport-related benefits,
energy-environment-economy modelling discussion paper no. 5. University of Cambridge, Cambridge
Bateman IJ, Carson RT, Day BH et al. (2002) Economic valuation with stated preferences techniques: a
manual. Edward Elgar, Cheltenham
Bergmann A, Hanley N, Wright R (2006) Valuing the attributes of renewable energy investments. Energy
Policy 34(9):1004–1014
Berk R, Fovell R (1999) Public perceptions of climate change: a ‘willingness to pay’ assessment. Clim Chang
41:413–446
Berrens RP, Bohara AK, Jenkins-Smith HC, Silva CL, Weimer DL (2004) Information and effort in contingent
valuation surveys: application to global climate change using national internet samples. J Environ Econ
Manag 47:331–363
Blomquist GC, Whitehead JC (1998) Resource quality information and validity of willingness to pay in con-
tingent valuation. Resour Energy Econ 20:179–196
BG (2008) Basque plan to Combat climate change 2008–2012. Basque Government
Burtraw D, Krupnick A, Palmer K, Paul A, Toman M, Bloyd C (2003) Ancillary benefits of reduced air pollu-
tion in the US from moderate greenhouse gas mitigation policies in the electricity sector. J Environ Econ
Manag 45:650–673
Carson RT, Hanemann WM (2005) Contingent valuation. In: Mäler KG, Vincent JR (eds) Handbook of
environmental econoimcs. Valuing environmental changes. Elsevier, Amsterdam pp 821–936
Chib S, Greenberg E (1998) Analysis of multivariate probit models. Biometrika 85:347–361
Complainville C, Martins JO (1994) NOX/SOX emissions and carbon abatement, OECD Working paper,
No. 151. Paris
Danielson L, Hoban TJ, Houtven GVan , Whitehead JC (1995) Measuring the benefits of local public goods:
environmental quality in Gaston county. North Carol Appl Econ 27:1253–1260
Dowlatabadi H, Morgan MG (1993) A model framework for integrated studies of the climate problem. Energy
Policy 21:209–221
Euskobarometro (2008) Estudio periódico de la opinión pública vasca. http://www.ehu.es/cpvweb/
euskobarometro. Accessed on 01.10.2008
EUSTAT (2008) Basque Institute of Statistics. Available at. http://www.eustat.es. Accessed on 01.10.2008
Freeman MA III (2003) The measurement of environmental and resource values: theory and methods. 2.
Reources for the Future, Washington 491
Fundación BBVA (2008) Percepciones y actitudes de los españoles hacia el calentamiento global. Available
at. http://www.fbbva.es. Accessed on 01.10.2008
Greene WH (2007) LIMDEP version 9.0 user’s manual. Econometric Software, New York
Hanemann WM (1984) Welfare evaluations in contingent valuation information with discrete responses. Am
J Agric Econ 66:332–341
Heintz RJ, Tol RSJ (1996) Secondary Bene¯ts of climate control policies: implications for the global environ-
mental facility; CSERGE working paper GEC 96–17. University of East Anglia, Norwich
Hidano N, Kato T (2007) Economic evaluation of anti-global warming policies: determining variability of
WTP values
Hoehn JP, Loomis JB (1993) Substitution effects in the valuation of multiple environmental programs.
J Environ Econ Manag 25:56–75
Hoyos D, Mariel P, Fernandez-Macho J (2009) The influence of cultural identity on the WTP to protect natural
resources: some empirical evidence. Ecol Econ 68:2372–2381
INE (2008) Spanish National Statistics Institute. http://www.ine.es. Accessed on 01.10.2008
123
140 A. Longo et al.
IPCC (2001) Climate change 2001. Mitigation. Cambridge University Press, Cambridge
IPCC (2007) Climate change 2007. Mitigation of climate change. Cambridge University Press, Cambridge
Krupnick A, Alberini A, Cropper ML, Simon N, O’Brien B, Goeree R, Heintzelman M (2002) Age, health, and
the willingness to pay for mortality risk reductions: a contingent valuation survey of Ontario residents.
J Risk Uncertain 24:161–186
Layton DF, Brown G (2000) Heterogeneous preferences regarding global climate change. Rev Econ Stat
82:616–624
Li H, Berrens RP, Bohara AK, Jenkins-Smith HC, Silva CL, Weimer DL (2004) Would developing country
commitments affect US households’ support for a modified Kyoto Protocol?. Ecol Econ 48:329–343
Li H, Berrens RP, Bohara AK, Jenkins-Smith HC, Silva CL, Weimer DL (2005) Testing for budget constraint
effects in a national advisory referendum survey on the Kyoto protocol. J Agric Resour Econ 30:350–366
Longo A, Markandya A, Petrucci M (2008) The internalization of externalities in the production of electricity:
willingness to pay for the attributes of a policy for renewable energy. Ecol Econ 67:140–152
Maddison D (1995) A cost-benefit analysis of slowing climate change. Energy Policy 23:337–346
Markandya A, Rübbelke DTG (2004) Ancillary benefits of climate policy. Jahrbücher für Nationalökonomie
Stat 224:488–503
Mitchell RC, Carson RT (1989) Using surveys to value public goods: the contingent valuation method. RFF
Press, Washington
Nomura N, Akai M (2004) Willingness to pay for green electricity in Japan as estimated through contingent
valuation method. Appl Energy 78:453–463
Nordhaus WD (1994) Managing the global commons: the economics of climate change. MIT Press, Cam-
bridge
OECD (2000) Ancillary benefits and costs of greenhouse gas mitigation. OECD, Washington
Olsthoorn X, Amann M, Bartonova A, Clench-Aas J, Cofala J, Dorland K, Guerreiro C, Hendriksen JF,
Jansen H, Larsen S (1999) Cost benefit analysis of European air quality targets for sulphur dioxide,
nitrogen dioxide, fine and suspended particulate matter in cities. Environ Resour Econ 14:333–351
Park T, Loomis J (1996) Joint estimation of contingent valuation survey responses. Environ Res Econ 7:
149–162
Payne JW, Schkade DA, Desvousges WH, Aultman C (2000) Valuation of multiple environmental programs.
J Risk Uncertain 21:95–115
Pearce D (2000) Policy framework for the ancillary benefits of climate change policies. Ancillary benefits and
costs of greenhouse gas mitigation. OECD, Washington 517–560
Pittel K, Rübbelke DTG (2008) Climate policy and ancillary benefits: a survey and integration into the mod-
elling of international negotiations on climate change. Ecol Econ 68:210–220
Poe GL, Welsh MP, Champ PA (1997) Measuring the difference in mean willingness to pay when dichotomous
choice contingent valuation responses are not independent. Land Econ 73:255–267
Popp D (2001) Altruism and the demand for environmental quality. Land Econ 77(3):339–349
Roe B, Teisl MF, Levy A, Russell M (2001) US consumers’ willingness to pay for green electricity. Energy
Policy 29:917–925
Rübbelke DTG (2002) International climate policy to combat global warming: an analysis of the ancillary
benefits of reducing carbon emissions. Edward Elgar Publishing Limited, Cheltenham
Tseng W, Chen C, Chang C, Chu Y (2009) Estimating the economic impact of climate change on infectious
diseases: a case study on dengue fever in Taiwan. Clim Chang 92:123–140
van den Bergh JCJM (2010) Safe climate policy is affordable—12 reasons. Clim Chang 101(3–4):339–385
Veisten K, Hoen HF, Navrud S, Strand J (2004) Scope insensitivity in contingent valuation o f complex envi-
ronmental amenities. J Environ Manag 73:317–331
Witzke HP, Urfei G (2001) Willingness to pay for environmental protection in Germany: coping with the
regional dimension. Reg Stud 35(3):207–214
123
