Environmental and Resource Costs Under Water ... - Springer Link

Water Resour Manage (2011) 25:1615–1633 DOI 10.1007/s11269-010-9764-z

Environmental and Resource Costs Under Water Scarcity Conditions: An Estimation in the Context of the European Water Framework Directive Julia Martin-Ortega · Giacomo Giannoccaro · Julio Berbel

Received: 19 February 2010 / Accepted: 6 December 2010 / Published online: 18 January 2011 © Springer Science+Business Media B.V. 2011

Abstract The European Water Framework Directive (WFD) requires the good ecological status of surface water bodies, which implies the improvement of both their physicochemical condition, as well as their flow and continuity. The WFD prescribes the assessment of environmental and resource costs and benefits associated with implementing these improvements. The recent literature focuses almost exclusively on the assessment of the economic values related to quality aspects. However, in much of southern Europe, fulfilling the WFD goals will greatly depend on maintaining sufficient water flow, as well. This study aims to fill this gap by assessing the non-market value of allocating enough water to the environment to ensure environmental services are sustained when water is scarce. The non-market value of guaranteeing water supply for secondary household uses is also estimated. Using the Guadalquivir River Basin in Spain as a case study, a choice experiment is applied with scenarios characterized by varying water flow levels and accompanying environmental impacts, and a different frequency of household water restrictions. The results show that the population derives significant benefits not only from the direct use of water, but that also holds non-use values related to the ecological status, although the latter has a considerably lower impact on consumer surplus. Additionally, we conclude that the costs of implementing the water saving measures

The research was done while Julia Martin-Ortega was affiliated to the University of Cordoba (Spain) and guest researcher at the Institute for Environmental Studies, VU University in Amsterdam (The Netherlands). J. Martin-Ortega (B) Catchment Management Group, Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen, AB15 8QH, UK e-mail: [email protected] G. Giannoccaro · J. Berbel Department of Agricultural Economics, University of Córdoba, Campus de Rabanales, Edificio C5 Tercera Planta, 14017, Córdoba, Spain

1616

J. Martin-Ortega et al.

currently included in the Program of Measures seem to be proportionate to its benefits in this case. Keywords Environmental and resource costs · Good ecological status · Choice experiment · Water Framework Directive · Water scarcity

1 Introduction The European Water Framework Directive (WFD) constitutes a major regulatory reform of water resource management within the European Union in which integrated catchment management plans must be prepared for all EU water bodies in order to achieve the ’good ecological status’. One of the key innovative elements of the WFD is the prescription of economic tools and principles for the achievement of this ecological objective. For instance, Member States have to take account of the principle of cost recovery of water services, including environmental and resource costs, i.e. in principle, water users must be charged with the full costs of water provision (article 9). Moreover, in case a derogation of the ecological objectives is sought on the basis of disproportionate costs, all costs and benefits need to be assessed; including environmental and resource costs and benefits (article 4). Up to now, the literature on the assessment of environmental and resource costs and benefits in the context of the WFD has been dominated by the estimation of the environmental values related to water quality. For example, in the valuation study carried out by Hanley et al. (2006), the environmental benefits of the implementation of the WFD in two British river basins are estimated. The authors base their valuation on three river quality attributes: in-stream ecology, aesthetics/appearance and bank-side conditions. Del Saz-Salazar et al. (2009), Bateman et al. (2009), Martin-Ortega and Berbel (2010), and Brouwer et al. (2010) use different versions of a water quality ladder in their valuations of environmental benefits in different European river basins. The water quality ladder is an instrument that describes open water quality in an ascending scale of water-use possibilities and was first used by Carson and Mitchell (1993) in the United States. As said, all these studies only address environmental costs and benefits of quality related aspects of the ecological status. In their definition of the ecological status, the volume of water that is actually allocated to the ‘environmental use’ is not considered. However, maintaining a certain level of water flow in the river is a pre-condition for the achievement of the good ecological status and the provision of environmental services, such as habitat protection. This is recognized in the Directive which defines ‘ecological status’ as ‘an expression of the quality of the structure and functioning of aquatic ecosystems associated with surface waters’ (article 2, item 21). This definition includes aspects related to the hydrological regime (i.e. the quantity and dynamics of flow), river continuity and morphological conditions affected by channel patterns (Annex V, Section 1.2.1). For a discussion on the indicators and definition of the ‘ecological status’ in the context of the WFD see Carballo et al. (2009). Quantitative aspects related to the ecological status are of particular importance in southern Europe (and the rest of the Mediterranean), where all indicators point to

Environmental and Resource Costs Under Water Scarcity Conditions

1617

an increase in water scarcity problems with negative implications towards current and future sustainability (Iglesias et al. 2007). The pressure on the resource is expected to increase due to predicted impacts of climate change that suggest a dryer and warmer Mediterranean region (e.g. Huntingford et al. 2003; IPCC 2007). With the resource under pressure, strong competition among different water users, including the environment, makes it necessary to estimate the opportunity costs of the possible allocation options. If extraction depletes the river flow to such an extent that the good ecological status cannot be maintained, the resulting loss in environmental value—or the forgone benefits of having a good ecological status— should also be considered among the opportunity costs. The opportunity costs of some water uses can be assessed through market prices. For example, the marginal value of water for irrigation can be estimated by using production functions (MesaJurado et al. 2010). Environmental value, on the other hand, involves direct, indirect and passive use values that are not reflected in the market. The forgone environmental benefits, as perceived by society, do need to be assessed as a pre-requisite for the maximization of social utility when making water allocation decisions. See Birol et al. (2006a, b) for a comprehensive taxonomy of the values composing the total economic value of water. The study presented here contributes to the recent literature by focusing on environmental values related to water quantity, and does this by estimating the nonmarket benefits of maintaining a sufficient stream flow for the good ecological status to be attained. These benefits equate to the opportunity costs of not leaving enough water in the river for environmental services to be provided. Besides looking at the environmental and resource costs, the study also estimates the value of household water availability. As is true for the environmental value, the value of having enough water in the household does not have a market price, since water prices do not reflect the real social value of water. It should be noted that ensuring water supply to secondary household uses, or reducing restrictions on these uses, is not an objective of the Directive, which focuses on the achievement of the good ecological status. Therefore, it could be argued that estimating the value of unrestricted household water supply for secondary uses is not strictly necessary for the resource cost assessment in the context of the Directive. However, because there is no market value for this water use and it is a matter of important social concern, it is a very policy relevant assessment. Therefore, indicators of the social value of providing water for household uses are also needed for a complete estimation of the non-market values of water use. The Guadalquivir River Basin in southern Spain was used as a case study to estimate the household willingness to pay (WTP) for the abovementioned benefits. A choice experiment was applied with scenarios characterized by different levels of the river’s water flow, causing varying degrees of deviation from natural ecological conditions with respect to species diversity and habitat, and a different frequency of household water restrictions. The remainder of this paper is organized as follows. In Section 2 the notion of resource costs in the context of the WFD is discussed. In Section 3 the methodology is presented. Section 4 describes the case study and the experimental design. Results are presented in Section 5 and conclusions are discussed in Section 6.

1618


2 Environmental and Resource Costs in the Water Framework Directive The Water Framework Directive prescribes the inclusion of environmental and resource costs of water use for the economic analysis of its implementation. It does not however, provide with a definition of these concepts and the current Programs of Measures of different Member States do not make a clear distinction between them. In the guiding document WATECO for the implementation of the economics aspects of the WFD (European Commission 2003) and various working groups under the Common Implementation Strategy, environmental and resource costs have been split up in two separate categories. In WATECO, environmental costs are defined as the costs derived from the environmental damage that water use imposes on ecosystems (e.g. the costs related to the decrease of water quality due to the pollution of an aquatic system). Resource costs, on the other hand, are specified to represent the costs of foregone opportunities due to the depletion of the resource beyond its natural rate of recharge or recovery. A later interpretation by the ECO2 working group put the emphasis in the definition of resource costs on the difference between the economic value of the current use of water and the economic value of the best alternative use. Therefore, according to ECO2, resource costs arise if another water use exists which would generate higher economic value than the current one. This definition is not restricted to the case of environmental damage or depletion. It could therefore be interpreted that, within the context of the WFD, environmental costs will arise when there is environmental damage, while resource costs refer to the opportunity costs of competing water uses. One should however be careful with this distinction, as opportunity costs can also arise in the case of environmental damage, for instance, if a water body gets polluted and is no longer available for another user. The interpretation can also be that environmental costs refer to the cost related to water quality, and resource costs to the cost related to water quantity issues. But, once again, this is not straightforward as water quantity and water quality are intrinsically interwoven. For example, a lower volume of water implies higher concentrations of pollutants. In this sense, it seems that environmental and resource costs are notions that are conceptually difficult to separate. In the context of the WFD, the environmental change that needs to be assessed is the difference between the baseline state of a water body or river basin and its good ecological status. What then really matters are the characteristics of the water management problem at hand, and whether water availability is a factor in the attainment of this good ecological status, as is the case in large parts of southern Europe. Therefore, environmental and resource costs are defined here together as the total economic value of the environmental damage as a result of the gap between the current and good ecological status of water bodies. This includes the economic value of the opportunities foregone of not reaching the WFD environmental objectives under scarcity conditions due to existing water allocation rules.1

1 This

is the definition adopted in the context of the technical guideliness for the assessment of environmental and resource costs of the WFD produced by the AquaMoney project in which context this research was carried out (Brouwer et al. 2009).


1619

3 Methodology: Choice Experiments Choice experiments are a stated preference technique for environmental valuation which involve eliciting responses from individuals in constructed, hypothetical markets. Choice experiments are rooted in Lancaster’s (1966) Value Theory and consistent with Thurstone’s Random Utility Theory (1929). They are frequently used in the valuation of non-market environmental goods (Hanley et al. 2001; Bateman et al. 2002). Choice experiments are said to have a number of advantages over contingent valuation (CV), including the provision of an ‘internal’ scope test (Hanley et al. 2001) and superior conditions for benefits transfer (Morrison et al. 2002). The method was chosen for this study because of its ability to deal with different attributes of one good simultaneously (Hanley et al. 1998). In the specific case of assessing the environmental values of water supply options, choice experiments have been found to be more flexible and cost-effective when several alternative proposals need to be considered (Blamey et al. 1999). In choice experiments, respondents in a survey are asked to choose between different bundles of (environmental) goods which are described in terms of their attributes and the levels of these attributes (Bennett and Blamey 2001). The economic theory underlying choice experiments assumes that the most preferred option yields the highest utility for the respondent (Louviere and Hensher 1983). One of the attributes is a monetary one, which makes it possible to estimate Hicksian compensating variation welfare measures for environmental changes. The most commonly applied structure of choice models is the Conditional Logit Model (CLM). The CLM assumes that the random components of the utility of the alternatives are independently and identically (i.i.d.) Gumbel distributed with a type I extreme value distribution. This assumption leads to a closed-form mathematical model that enables estimation by conventional maximum likelihood procedures. The standard indirect utility function underlying the CLM is: U ij = Vij + εij = βk Xij + εij

(1)

where U ij refers to the utility of household i obtained from choice alternative j; Vij is the measurable component of utility, measured through a vector of k utility coefficients β associated with a vector of attribute and individual characteristics Xij; and εij captures the unobserved influences on a household’s choice (error term). The CLM can be restrictive in practice. The Random Parameters Logit (RPL) or Mixed Logit model is more flexible and relaxes the assumption of independence of irrelevant alternatives (IIA) that results from the i.i.d. property underlying the CLM, and therefore allows for preference heterogeneity (McFadden and Train 2000; Train 2003). Other models also exist, such as the Nested Model and the Latent Class Model. See Hensher et al. (2005a) and Train (2003) for detailed information on the different kinds of choice models. 4 Case Study Description and Experimental Design 4.1 The Guadalquivir River Basin The Guadalquivir River Basin (GRB) is the longest river basin in southern Spain, with a length of around 650 km. It covers an area of 57,527 km2 with a population

1620


of over 4 million inhabitants. The basin has a Mediterranean climate with a heterogeneous precipitation distribution (the annual average temperature is 16.8◦ C, and annual average precipitation is 630 mm). The natural surface water resources of the GRB yield up to 7,230 Hm3 /year, from which only 40% is available for use. Groundwater yields 2,576 Hm3 /year, from which 53% is currently used (Confederación Hidrográfica del Guadalquivir 2010a). Currently, the main water uses in the GRB are agriculture (87%), domestic use (9%), industrial use (3%) and tourism (1%) (Martin-Ortega et al. 2008). The pressure on the resource is expected to rise in the coming years: water demand is expected to increase, with consumption surpassing 50% of the renewable flow by 2015 (Gutiérrez et al. 2009). At the same time, climate change, through both a rise in temperatures and a reduction in rainfall, will result in lower water yields and further raise demand for irrigation. For the horizon 2030, simulations considering a temperature increase of 1◦ C and a reduction of 5% in mean rainfall project a decrease of mean hydraulic yields of almost 12% in the GRB, above the national Spanish average of 8% (Iglesias et al. 2005). The Spanish implementation of the WFD imposes the maintenance of a minimum water flow in the river basin, which would imply limitations on the consumptive use of water in other sectors. If WFD prescriptions are respected, the current water use should already be reduced by 6.5% (Confederación Hidrográfica del Guadalquivir 2008). In the context of expected demand increase and precipitation decrease, the conflict between competing water uses and the maintenance of the ecological status is therefore guaranteed. 4.2 Valuation Design Respondents from a representative sample of the river basin’s population were asked to select the most preferred alternative in a set of different scenarios of water allocation to the environment and frequencies of household water restrictions. The preparation for the valuation design included a literature review, discussions within the AquaMoney Project Water Scarcity Group and expert consultation including its presentation and discussion in several scientific workshops organized within the Project (www.aquamoney.org). Also, two focus groups were organized, in which in total 30 people participated and three rounds of pre-testing were held, involving 110 individuals. The three pre-tests focused on: (1) the selection of attributes and levels, (2) the testing for the clarity and credibility of the scenario and the baseline situation, and (3) fine tuning of all questions. Regarding the water allocation to the environment, respondents were informed that securing future water supply for the environment meant that water levels in rivers and aquifers increase, reducing both the risk that they will dry out irreversibly, while also improving living conditions for a variety of fish, plants and other animals living in and near the water. Three possible levels of improvement were proposed (moderate, good and very good), corresponding to the WFD categorization, and compared to the expected poor situation with low water levels. The descriptions of the levels in the survey match the definitions given in the Directive (Section 1.2.1, Annex V), but the labelling of the levels is different. The level ‘very good’ in the survey scale corresponds to the ‘good ecological status’ of the Directive. Pre-testing


1621

showed that the labels chosen for the survey were easier to understand. See Table 1 for a detailed description of the attribute levels. The estimation of the willingness to pay for higher water supply standards for household use is not new in the literature. Often water quality and quantity are assessed together. For example, Genius et al. (2008), Genius and Tsagarakis (2006) and Vásques et al. (2009) evaluate consumer’s WTP for financing projects to improve potable water quality and quantity in Greece and Mexico, respectively. Haider and Rasid (2002) assess people’s preferences regarding water rates, water pressure and water taste from different water sources in Canada. Raje et al. (2002) apply a CV study to elicit values for improved water supply conditions in India. Other studies relate the WTP for water supply to urban wastewater management (Kontogianni

Table 1 Description of the choice experiment attributes and levels Attributes

Description

Attribute 1: Levels of ecological status related to water flow Poor (baseline)

This is the expected future situation of low water levels and environmental quality status. It entails a large deviation from the natural situation due to the increased water scarcity conditions. Many fish species will disappear and riverbanks will lose most of their vegetation. As a result most birds and insects will disappear, as well Moderate Less than natural water levels and environmental quality status. This is still a substantial deviation from the natural situation. A limited number of fish species are present. Riverbanks have some vegetation supporting a limited number and variety of birds and other wildlife Good Water levels and environmental quality status are close to their natural levels. There is a small deviation from the natural situation. Under these conditions riverbanks have a lower than natural vegetation cover. As a result the breeding and nesting conditions for a number of species are limited Very good Water levels and the environmental quality status are at their natural state where there is almost no pressure from human activities. Conditions for wildlife are optimal under these circumstances Attribute 2: Frequency of household water restrictions Restrictions in 4 out of the next 10 years (baseline) Restrictions in 3 out of the next 10 years Restrictions in 2 out of the next 10 years Restrictions in 1 out of the next 10 years Monetary attribute

Restrictions on watering gardens, filling swimming pools, using washing machines and other ‘secondary water uses’. Water shortfall lasting 6 to 7 hours during up to 20 days in dry summers Increase in the household water bill of e0 (baseline) 20, 40, 60, 80, 100 and 120 per year during the next 10 years

1622


et al. 2003; Xenarios and Bithas 2009; Alcon et al. 2010). In our study, we specifically wanted to assess the value of avoided water restrictions. Several approaches were tested. We first tested water restrictions expressed as the probability (in percentages) that certain water shortages of a given frequency and severity would occur. This was done in a similar way as Koss and Khawaja (2001), who apply the CV method to a case study in California. This approach proved to be to complex, as some respondents in the pre-test did not understand the probabilities correctly. We then opted to present water restrictions in terms of the number of years in which restrictions were expected to occur within the next 10 years. This follows the approach of Hensher et al. (2005b) and Hatton McDonald et al. (2010), who apply a choice experiment to obtain implicit prices for the reliability of household water services in Australia. Among other attributes, those authors include the frequency of water interruptions, ranging from ‘12 times per year’ to ‘once every 10 years’ in the study by Hensher et al. (2005b) and from ‘no more times’ to ‘two more times’ in the next 12 months in the case of Hatton McDonald et al. (2010). In our study, the restrictions consisted of households not being allowed to water gardens, fill swimming pools, use washing machines or enjoy other ‘secondary uses’ of water during certain hours in dry summers. The base level was that restrictions occur in 4 out of the next 10 years. Potential improvements reduce this to occurrences in 3, 2 or 1 years out of the next 10. Respondents were informed that the water shortfall could last up to 20 days with water restrictions in place for up to 6 to 7 hours a day. At early stages of the investigation, the possibility of including full water restrictions (i.e. also for drinking and personal hygiene) was contemplated. However, this generated many protest responses as water supply for primary household use is prioritised by law. To our knowledge our study is the only one combining the assessment of both household water supply reliability and environmental water use, and specific to the WFD. Two different approaches can be taken to construct experimental designs for CEs. The first one is based on probability balanced designs, and does not use any prior information about parameters. The second is based on increasing design efficiency by making assumptions about the sign or relative size of the parameters (Scarpa and Rose 2008). The first design approach has been widely adopted in the literature (e.g. Blamey et al. 2000; Haider and Rasid 2002; Birol et al. 2006a, b; Karousakis and Birol 2008; Rolfe and Bennett 2009; Martin-Ortega and Berbel 2010), and is the one used in this study. It results in orthogonal experimental designs, where attributes are uncorrelated (Carlsson and Martinsson 2003). It should be noted that several authors have recently proposed that orthogonal designs may not be the most efficient when complex non-linear models are used (Ferrini and Scarpa 2007; Rose and Bliemer 2008; Scarpa and Rose 2008). However, with large enough samples, orthogonal designs still lead to reliable results (Scarpa and Rose 2008). The fractional factorial design used here is composed of 24 choice cards, organized in 6 choice sets. Each responded faced four choice occasions in which they were asked to choose between the current situation with no rise in the water bill, and option A and B with different levels of improvement of the two attributes in exchange for an annual increase in the water bill. The increase in the water bill was identified in the pre-test phase as the best payment vehicle. Moreover, it has been successfully applied in water resource valuation studies in the past (see for example, Genius et al. (2008); Martin-Ortega and Berbel 2010; Del Saz-Salazar et al. 2009, Brouwer et al. 2010;


1623

Fig. 1 Choice card example (translated from Spanish). The pictograms included in the choice cards were elaborated within the Water Scarcity Group of the AquaMoney Project

Alcon et al. 2010). Baker et al. (2007) also use the water bill as a payment vehicle, but in combination with higher prices on everyday products. The monetary attribute included six different levels: no increase (current water bill), and increases of e20, e40, e60, e80, e100 and e120 in the annual household water bill during the next 10 years. The levels of the monetary attribute were determined through a pre-test with an open-ended WTP question. The average current water bill in the region is e247 per household per year (Spanish Association for Water Supply and Water Treatment AEAS 2008). Figure 1 shows an example of a choice card. 4.3 Questionnaire Structure and Sampling Procedure The questionnaire was organized in three sections. The first section was aimed at capturing respondents’ perceptions of the environment and water scarcity, and their prior experience with the latter. For example, respondents were asked ‘Do you think water scarcity is a problem in your region?’, ‘Have you suffered water restrictions in your household in the past?’, ‘Do you recreate in the river basin?’, and ‘Do you think the ecological status of the river basin is affected by the lack of water?’. The second section contained the valuation scenario with the four different choice occasions. The actual choices were preceded by questions about the credibility of and acquaintance with the provided information about the status quo. Intense pre-testing at several stages resulted in a satisfactory level of credibility of the proposed scenario: A large majority of the respondents considered the presented scenario credible (87% found it credible, and 96% found it credible or somewhat credible). Three quarters

1624 Table 2 Main sociodemographic characteristics of the sample and population

a Source

of target population data: Regional Institute for Statistics (http://www. juntadeandalucia.es/ institutodeestadistica/). Year 2008. Cited Feb 2010

J. Martin-Ortega et al. Variables Age groups (%) 18–34 years 35–59 years ≥60 years Gender (% women) Household size (# of people) Household with children (% of total) Monthly household income (e) Level of education (% of total) None Lower secondary Upper secondary education

Sample

Target populationa

31.4 51.1 17.5 50.3 3.5 50.8 1,874

32.5 48.8 18.7 51.4 2.8 63.8 1,763

12.7 58.7 25.7

17.1 69.9 13

of the sample thought it likely that water restrictions will occur in the future and 22% believed that they will definitely occur. Demographic and socio-economic information of the respondents was gathered in the final section of the questionnaire. A total of 354 respondents were interviewed following a stratified sampling procedure with sex and age quotas. The survey took place in nine rural and urban municipalities of the river basin during the months of July and September 2008. The main demographic and socio-economic characteristics of the sample are presented in Table 2. Overall, the sample is a fairly good representation of the population of the Guadalquivir River Basin.

5 Results 5.1 Sample Characteristics and Perceptions When asked about the most important problems in the region, 32 respondents (9% of the sample) mentioned water as one of them. The most frequently mentioned problems are related to the economy (economic crisis, unemployment, inflation, etc.) and to urban life and insecurity (e.g. traffic and criminality). When asked directly, more than 94% of the sample stated that environment in general is important or very important, although only 1% belongs to an environmental organization. Eighty four percent thinks that water availability in the GRB is a problem or a great problem (39.5 and 44.6%, respectively). The majority of the sample (85%) thinks that water scarcity in the GRB will increase in the future. Moreover, 322 respondents (91%) believe that the environment is affected by water availability in one way or another. About a third of the sample considers that water scarcity is a natural phenomenon beyond human control, while 50% does not agree with this statement. Approximately 60% of the sample has suffered water restrictions in the household. Only 4% remembers to have seen social awareness campaigns for water use savings. A majority of the sample (63%) believes that they will, either probably or without a doubt, suffer water restrictions in their household in the future. Half of the sample believes that these restrictions will occur in 1 or 2 out of 10 years.


1625

The recreational use of the river is very low. Only 4% recreates in the river. A similar percentage (5%) considers that their work is affected by the lack of water, most of them farmers. These figures are similar to what was found in previous studies in the same region (Mesa et al. 2008; Martin-Ortega et al. 2009). Respondents were asked if they remembered the amount on their latest water bill. Almost half of the sample (47.7%) did. The average of these amounts was e174.04 per household per year. It should be noted that this ‘remembered’ amount is significantly lower than the actual average amount that residents pay, which is e247 per household per year according to the Spanish Association for Water Supply and Water Treatment AEAS (2008). 5.2 Monetary Values Protests answers were distinguished from legitimate zero bids by means of follow up questions about the reasons respondents were not willing to pay. Legitimate zero bids correspond to individuals who attach no value to the services under assessment. In our study legitimate zero answers include respondents who ‘do not think the good is important’, ‘cannot afford to pay extra’, ‘prefer to spend the money on other things’, or ‘feel they pay enough already’. ‘It is a problem of the State’, was considered a protest against the payment vehicle. Less than 5% of the responses were classified as protests and were excluded from the WTP analysis, as is the common practice in the literature (Whitehead et al. 1993; Mitchell and Carson 1989; Jorgensen et al. 1999; Dziegielewska and Mendelsohn 2007). Average WTP values were calculated using the Conditional Logit Model. The attribute ‘reduction in the frequency of household water restrictions’ was coded as a continuous variable. This coding allowed us to estimate the value of reducing the probability of water restrictions by one year out of ten, within the range of years used in the survey (1–4). The environmental attribute was coded as a set of dummy variables, one for each of the levels of water flow improvement. For example, the variable ‘very good ecological status’ takes the value 1 if the choice alternative represents an improvement from the current status to the very good level, and zero otherwise. In the model, we also include the cost attribute and an alternative specific constant (ASC) which takes value 1 when either scenario A or B is selected and zero when the current expectation scenario is selected. Table 3 presents the modelling results. The model is statistically significant, with all attributes being highly significant and having the expected sign. The attribute ‘frequency of restrictions’ has a negative sign, implying that households derive utility from reducing the frequency of restrictions. The environmental dummies show increasing utility for increased levels of improvement (i.e. the utility derived by improving the environmental status from low to very good is higher than from low to good). This proofs, as theoretically expected, that respondents are sensitive to scope (Carson 1997). The monetary attribute is also statistically significant and negative, implying that respondents are price sensitive (i.e. the higher the costs, the lower the utility of the choice option). The positive significant sign of the ASC coefficient implies that a positive utility is associated with a move away from the status quo.

1626


Table 3 Conditional Logit model of choice attributes

a Significant

at the 1% level

Variable ASC Frequency of household water restrictions Good ecological status Very good ecological status Cost Num. observations Log likelihood R2

Coefficient

Std. error

2.37a −1.65a

0.20 0.55

0.33a 0.45a −0.04a 1,348 −1,227.48 0.16

0.10 0.11 0.01

As discussed in Section 3, the CLM is based on the assumption of the irrelevance of independent alternatives (IIA) (Train 2003). To test whether this assumption holds, the Hausman test, based on the exclusion of one of the alternatives, can be applied (Hausman and McFadden 1984). The test resulted in a non significant difference (χ 2 = 3.167, Pr (C > c) = 0.674), implying that the IIA assumption is not violated in this case (Hensher et al. 2005a).2 Average WTP for each attribute is derived from the model based on expression (2) (Train 2003). PIk = − (βk /βc )

(2)

where PIk is the implicit price (WTP) for attribute k; β k is the coefficient of the attribute k; and β c is the coefficient of the cost attribute. Results including standard deviations calculated using the Delta Method (Greene 2003) are presented in Table 4. The average WTP for a reduction in the frequency of household water restrictions by 1 year out of the next ten is e39.53 per household annually. Average WTP for the allocation of water to the river basin for the maintenance of environmental services amounts to e7.95 and e10.88 per household per year, for an improvement to a good and a very good ecological state, respectively. These results imply that households are not only willing to pay to secure their own water in case of scarcity, but also hold non-use values for the maintenance of the ecological status of the river and the ecosystem services it provides. It could be argued that the value attached to the ecological status relates to the recreational use value, but recreation was modelled in the multivariate analysis (see Section 5.4) and found to be statistically non-significant. This means that we can assume that an important share of the value is non-use value.

2A

Random Parameter Logit model was estimated, nevertheless. It was based on 500 Halton simulations and allowed for a normal distribution of the coefficient of the ‘frequency of water restrictions’ attribute and a uniform distribution of the coefficient of the ecological status attributes, as recommended for dummy variables by Hensher et al. (2005a). In the model, the standard deviation of most of the random coefficients is not significant, and it has a very similar overall fit as the CLM (results available upon request). It is therefore concluded that the RPL model does not improve the explanatory power.


1627

Table 4 Average WTP for decreased frequency of water restrictions and improvement of the ecological status in e per household per year Attribute

Average WTP

Std. error

Reduction in frequency of household water restrictions Good ecological status Very good ecological status

39.53 7.95 10.88

12.89 2.38 2.60

5.3 Water Policy Scenarios Choice modelling allows for the specification of policy scenarios with different levels of improvement, and the comparison of the non-market benefits of these scenarios. Table 5 presents three policy scenarios for which the non-market benefits have been calculated. In the first scenario, measures are taken to reduce household water restrictions to the minimum level, occurring only in 1 out of the next 10 years, while the ecological status of the river basin remains at the current low level. In the second scenario, the situation is opposite: the frequency of restrictions is maintained as in the status quo, occurring in 4 out of the next 10 years, but the ecological status improves to the very good level. Finally, the third scenario represents the maximum improvement of both attributes. The benefits per household of these scenarios are e119, e11, and e129, respectively. Table 5 also includes the percentage that these WTP values represent over the average water bill of e247 per household per year. It can be observed in Table 5 that, for scenario 3, which corresponds to the best scenario possible according to our design, inhabitants of the river basin are willing to increase their current water bill by more than 50%. If we compare the WTP values with the amount of money respondents thought they pay for water (e174.04 per household per year), instead of with the actual water bill, the WTP value represent a 74.4% increase. This implies that the population derives significant benefits from these improvements. 5.4 Demand Heterogeneity Another useful application of stated preference techniques is that they allow for the inclusion of household characteristics in the analysis, which can provide information on demand heterogeneity. Similarly to Genius et al. (2008), we use some of the

Table 5 Non-market benefits of water policy scenarios Scenario

Frequency of household water restrictions

Ecological status

Consumer surplus (e /household per year)

Std. error

Increase over average current water bill (%)a

1

1 year out of the next 10 4 years out of the next 10 (baseline) 1 year out of the next 10

Low (baseline)

118.60

12.14

48.02

Very good

10.88

5.30

4.40

Very good

129.48

11.77

52.42

2 3 a Source:

Percentages calculated over the annual water bill reported by the Spanish Association for Water Supply and Water Treatment AEAS (2008)

1628


individuals’ information gathered in the survey to explore the factors influencing WTP. Analyzing demand heterogeneity is possible in choice experiments via the interaction of respondents’ characteristics with the constant of the model and/or with the attributes (Rolfe et al. 2000). Two interactions models have been estimated, and are presented in Table 6. In the first one, the effect of income on WTP is investigated as theory tells us that households with higher incomes should be willing to pay more. Income is therefore interacted with the attributes representing the frequency of restrictions and the improvement of the ecological status to the highest level. In the second interaction model, we are interested in the effect that perception about water scarcity has in WTP. For this purpose and after several tests, we interact: (1) the variable ‘water scarcity is a problem in the region’ with the ASC of the model, (2) the variable ‘I don’t believe there will be water restrictions in my household in the future’ with the ‘frequency of restrictions’ attribute, and finally, (3) the variable ‘I think water allocation to the environment should receive priority in case of water scarcity’ with the dummy variable for the highest level of the ecological status attribute. It can be observed from the signs of the coefficients of the interaction terms in model 1 that the results confirm the expected effect of income on the WTP for both attributes.

Table 6 Conditional Logit model with interactions Variables

Interaction model 1: income effect

Attributes ASC Frequency of water restrictions Good ecological status Very good ecological status Cost Income * Frequency of restrictions Income * Very good ecological status Considers water scarcity a problem in the region * ASC Assumes to suffer water restrictions in the future * Frequency of water restrictions Believes water allocation to the environment should have priority in case of scarcity * Very good ecological status Num. observationsd Log likelihood R2 a Significant

Coef.

Std. error

Interaction model 2: effect of water scarcity perception Coef. Std. error

2.36c 0.51 0.34c −0.47 −0.04c

0.21 1.14 0.10 0.32 2*10−3

2.29c −1.32b 0.12 0.46c −0.04c

0.21 0.62 0.13 0.11 2*10−3

−0.96b 0.41b

4*10−5 1*10−4

– –

– –

– –

– –

0.47b −0.69a

0.23 0.52

–

–

0.45b

0.15

1,348 −1,222.38 16.2%

1,416 −1,196.42 18.2%

at the 10% level, b significant at the 5% level, c significant at the 1% level

d Observations

equal to 354 respondents * 4 choice occasions per respondents (minus missing values)


1629

From the interaction model 2 we obtain the following results: the interaction of the variable ‘I think water scarcity is a problem in the region’ with the ASC is statistically significant and positive, indicating that households are willing to pay more for a move from the current situation if they are concerned about water scarcity problems. The expectations about suffering water restrictions in the future also influence WTP: people who do not think that they will suffer water restrictions in their household are willing to pay less. Finally, people who gave preference to the environment when asked to prioritise different sectors for water distribution in case of water scarcity, not surprisingly, have a higher WTP than respondents who selected the agricultural or industrial sector. 6 Conclusions The good ecological status of all water bodies, as prescribed by the European Water Framework Directive, implies the improvement of both the physicochemical condition of the water, as well as its quantitative conditions (water flow and continuity). The value given by society to the allocation of enough water to the environment to ensure that environmental services are sustained when there is a water shortage has been scantily studied in the context of the WFD. The existing studies have focused almost exclusively on the values related to water quality. However, the depletion of river basins is a major issue for policy makers in southern Europe who are concerned with attaining the WFD’s good ecological status. The expected effects of climate change for the region will further increase the importance of the issue. This study aims to fill this gap in the literature by assessing the non-market costs (foregone benefits) of not allocating enough water to river basins to reach the good ecological status. The non-market value of reducing water restrictions for secondary household uses was also addressed. The case study for this research was the Guadalquivir River Basin, in southern Spain, in which the expected increase in the demand for water and the expected decrease of precipitation will imply enhanced competition for the resource. The applied choice experiment showed that the river basin population is concerned about water scarcity problems and is aware of the implications of the lack of water for the ecological status of the river basin, and that they are willing to pay to address these problems. They are willing to increase over 50% their current water bill (and almost 75% of their ‘remembered’ water bill). A first conclusion is that, when water is scarce, people not only have direct use values for the water, such as those related to domestic supply, but also hold non-use values associated with ecosystem services, such as habitat support. However, the impact of the ecological status of the river on consumer surplus is considerably lower than that of the restrictions of household water supply. The second conclusion relates to the proportion of the benefits to the costs of implementing river basin improvements. The Program of Measures prescribed by the WFD is currently being prepared for the Guadalquivir River Basin. The latest version thereof, dating from July 2010, includes investments with annual costs of e465 million for the period 2009–2015 (Confederación Hidrográfica del Guadalquivir 2010b). From these annual costs, 44%, or e205 million, correspond to water saving measures. This implies an equivalent annual cost of e119 per household (Berbel et al. 2010). Our results show that, for the best policy scenario (i.e.

1630


lowest level of water restrictions in households and highest environmental quality), total non-market benefits exceed these costs (e129.48). Therefore, the costs of implementing the water saving measures included in the Program of Measures seem to be proportionate to the benefits in this case and more so if the market benefits were added as well. Previous studies for the river basin focusing exclusively on quality improvements similarly show that the non-market benefits of these improvements compensate for the costs of the measures that relate to improving quality (Martin-Ortega 2010; Brouwer et al. 2010). However, the values from these different studies cannot simply be added up to the ones presented here, as quality and quantity issues are intrinsically interwoven, and this could lead to double counting of benefits. A full integration of both quantitative and qualitative perceptions and interactions for the comparison of the full costs of the Program of Measures will require further development and should be the next step in the research of perceived benefits of the WFD implementation. Acknowledgements This research was carried out in the context of the Collaboration Agreement between the University of Córdoba (Spain) and the Spanish Ministry of the Environment for the Development of Water Demand Analysis and Assessment of Environmental and Resource Benefits, and the AquaMoney project of the EU VI Framework Programme (SSPI-022723, Development and Testing of Guidelines for the Assessment of Environmental and Resource Costs and Benefits). The work was partly funded by the Spanish Ministry of Science and Innovation through the AGUAICAD project (ECO2009-12496-C03-01). The questionnaire was developed in collaboration with the members of the Water Scarcity Group of the AquaMoney project (Roy Brouwer, Julia MartinOrtega, Meri Raggi, David Viaggi, Manuel Pulido- Velázquez, Michalis Skourtos, Joaquín Andreu, Julio Berbel, Laura Sardonini, Davide Ronchi, Areti Kontogianni, Thanassi Machleras and Antonio Lopez-Nicolas), Begoña Alvarez-Farizo and Josefina Maestu. We are thankful to Sebastiaan Hess for his valuable comments. An earlier version of this work was presented as a working document (N◦ 497/2010) to the Fundación de las Cajas de Ahorros (FUNCAS). Only the authors are responsible for the results and statements in this article.

References Alcon F, Pedrero F, Martin-Ortega J, Arcas N, Alarcon JJ, de Miguel M (2010) The non-market value of reclaimed waste-water for use in agriculture. Span J Agric Res 8(S2) Baker B, Metcalfe P, Butler S, Gueron Y, Sheldon R, East J (2007) Report on the benefits of Water Framework Directive programmes of measures in England and Wales. DEFRA, United Kingdom Bateman IJ, Carson RT, Day B, Hanemann M, Hanley N, Hett T, Jones-Lee M, Loomes G, Mourato S, Özdemiroglu E, Pearce DW, Sugden R, Swanson J (2002) Economic valuation with stated preference techniques a manual. Edward Elgar, Cheltenham Bateman IJ, Brouwer R, Ferrini S, Schaafsma M, Barton DN, Dubgaard A, Hasler B, Hime S, Liekens I, Navrud S, De Nocker L, Šceponaviciut R, Semnien D (2009) Making benefit transfers work: transferring theoretically derived, spatially explicit, value functions for the non-market benefits of water quality improvements across Europe. CSERGE working papers EDM-2009–10. Available at: http://www.uea.ac.uk/env/cserge/pub/wp/edm/edm_2009_10.htm. Cited Sept 2010 Bennett JW, Blamey RK (2001) The choice modelling approach to environmental valuation. Edward Elgar, Cheltenham Berbel J, Martin–Ortega J, Mesa P (2010) A cost-effectiveness analysis of water-saving measures for the Water Framework Directive: the case of the Guadalquivir River Basin in Southern Spain. Water Resour Manage. doi:10.1007/s11269-010-9717-6 Birol E, Karousakis K, Koundouri P (2006a) Using economic valuation techniques to inform water resources management: a survey and critical appraisal of available techniques and an application. Sci Total Environ 365(1–3):105–122


1631

Birol E, Karousakis K, Kounduri P (2006b) Using a choice experiment to account for preference heterogeneity in wetland attributes: the case of Cheimaditida wetland in Greece. Ecol Econ 60(1):145–156 Blamey R, Gordon J, Chapman R (1999) Choice modelling: assessing the environmental values of water supply options. Aust J Agr Resour Ec 43(3):337–357 Blamey RK, Bennett JW, Louviere JJ, Morrison MD, Rolfe J (2000) A test of policy labels in environmental choice modelling studies. Ecol Econ 32(2):269–286 Brouwer R, Barton D, Bateman IJ, Brander L, Georgiu S, Martin-Ortega J, Navrud S, PulidoVelazquez M, Schaafsma M, Wagtendonk A (2009) Economic valuation of environmental and resource costs and benefits in the WFD: technical guidelines for practitioners. Available at: www.aquamoney.org Brouwer R, Martin-Ortega J, Berbel J (2010) Spatial heterogeneity in choice experiments. Land Econ 3(86):552–568 Carballo R, Cancela JJ, Iglesias G, Marin A, Neira XX, Cuesta TS (2009) WFD indicators and definition of the ecological status of rivers. Water Resour Manage 23:2231–2247 Carlsson F, Martinsson P (2003) Design techniques for stated preference methods in health economics. Health Econ 12:281–294 Carson RT (1997) Contingent valuation surveys and tests of insensitivity to scope. In: Kopp RJ, Pommerhene W, Schwartz N (eds) Determining the value of non-marketed goods: economic, psychological and policy relevant aspects of contingent valuation methods. Kluwer, Boston, pp 127–163 Carson RT, Mitchell R (1993) The value of clean water: the public’s willingness to pay for boatable, fishable and swimmable quality water. Water Resour Res 29(7):24–45 Confederación Hidrográfica del Guadalquivir (2008) Esquema de Temas Importantes de la Demarcación. Available at: http://www.chguadalquivir.es/opencms/portalchg/planHidrologicoDemarcacion/ participacionPublica/consultaPublica/#apartado1. Cited Sept 2010 Confederación Hidrográfica del Guadalquivir (2010a) La Demarcación. www.chguadalquivir.es. Cited Dec 2010 Confederación Hidrográfica del Guadalquivir (2010b) Guadalquivir River basin management plan. Discussion document for draft program of measures annex 10 (version July 2010) Dziegielewska DA, Mendelsohn R (2007) Does “No” mean “No”? A protest methodology. Environ Resour Econ 38:71–87 Del Saz-Salazar S, Hernández-Sancho F, Sala-Garrido R (2009) The social benefits of restoring water quality in the context of the Water Framework Directive: a comparison of willingness to pay and willingness to accept. Sci Total Environ 407(16):4574–4583 European Commission (2003) Economics and the environment: the implementation challenge of the 21 Water Framework Directive, WATECO: a guidance document. Brussels Ferrini S, Scarpa R (2007) Designs with a priori information for nonmarket valuation with choice experiments: a Monte Carlo study. J Environ Econ Manag 53:342–363 Genius M, Tsagarakis KP (2006) Water shortages and implied water quality: a contingent valuation study. Water Resour Res 42:W12407 Genius M, Hatzaki E, Kouromichelaki EM, Kouvakis G, Nikiforaki S, Tsagarakis KP (2008) Evaluating consumer’s willingness to pay for improved potable water quality and quantity. Water Res Manag 22:1825–1834 Greene WH (2003) Econometric analysis. Prentice Hall, New Jersey Gutiérrez C, Martin-Ortega J, Berbel J (2009) Situación y tendencias del uso agrícola del agua en la cuenca del Guadalquivir. Revista de Estudios Agrosociales y Pesqueros 220(4):163–176 Haider W, Rasid H (2002) Eliciting public preferences for municipal water supply options. Environ Impact Asses 22:337–360 Hanley N, Wright R, Adamowicz V (1998) Using choice experiments to value the environment. Environ Resour Econ 11(3–4):413–428 Hanley N, Mourato S, Wright R (2001) Choice modelling approaches: a superior alternative for environmental valuation? J Econ Surv 15(3):435–462 Hanley N, Wright R, Álvarez-Farizo B (2006) Estimating the economic value of improvements in river ecology using choice experiments: an application to the Water Framework Directive. J Environ Manag 78:183–193 Hatton McDonald D, Morrison MD, Barnes MB (2010) Willingness to pay and willingness to accept compensation for changes in urban water customer standards. Water Resour Manage 24:3145– 3158 Hausman J, McFadden D (1984) Specification tests for the multinomial logit model. Econometrica 52(5):1219–1240

1632


Hensher D, Rose J, Greene W (2005a) Applied choice analysis: a primer. Cambridge University Press, UK Hensher D, Shore N, Train K (2005b) Households’ willingness to pay for water service attributes. Environ Resource Econ 32(4):509–531 Huntingford C, Jones RG, Prudhomme C, Lamb R, Gash JHC, Jones DA (2003) Regional climatemodel predictions of extreme rainfall for a changing climate. Q J Roy Meteor Soc 129:1607–1621 Iglesias A, Estrela T, Gallart F (2005) Impactos sobre los recursos hídricos. In: Moreno J (ed) Evaluación preliminar de los impactos en España por el efecto del cambio climático. Ministerio de Medio Ambiente, Madrid, pp 303–352 Iglesias A, Garrote L, Flores F, Moneo M (2007) Challenges to manage the risk of water scarcity and climate change in the Mediterranean. Water Resour Manage 21:775–788 IPCC (2007) Working group I fourth assessment report the physical science basis. Geneva Jorgensen BS, Syme GJ, Bishop BJ, Nancarrow BE (1999) Protest responses in contingent valuation. Environ Resour Econ 14(1):131–150 Karousakis K, Birol E (2008) Investigating household preferences for kerbside recycling services in London: a choice experiment approach. J Environ Manage 88(4):1099–1108 Kontogianni A, Langford IH, Papandreou A, Skourtos MS (2003) Social preferences for improving water quality: an economic analysis of benefits from wastewater treatment. Water Resour Manage 17:317–336 Koss P, Khawaja S (2001) The value of water supply reliability in California: a contingent valuation study. Water Policy 3(2):165–174 Lancaster KJ (1966) A new approach to consumer theory. J Polit Econ 74(2):132–157 Louviere J, Hensher DA (1983) Using discrete choice models with experimental design data to forecast consumer demand for a unique cultural event. J Consum Res 10(3):348–361 Martin-Ortega J (2010) Los beneficios ambientales de las aguas del Guadalquivir: un análisis económico. Fundación Pública Centro de Estudios Andaluces ISBN: 978-84-692-9577-9. Available at: http://www.centrodeestudiosandaluces.es/. Cited Sept 2010 Martin-Ortega J, Berbel J (2010) Using multi-criteria analysis to explore non-market monetary values of water quality changes in the context of the Water Framework Directive. Sci Total Environ 408:3990–3997 Martin-Ortega J, Gutierrez C, Berbel J (2008) Caracterización de los usos del agua en el Guadalquivir en aplicacíon de la Directiva Marco del Agua. Rev Estud Reg 81:45–76 Martin-Ortega J, Berbel J, Brouwer R (2009) Valoración económica de los beneficios ambientales no de mercado derivados de la mejora de la calidad del agua: una estimación en aplicación de la Directiva Marco del Agua al Guadalquivir. Revista de Economía Agraria y de los Recursos Naturales 9(1):65–89 McFadden D, Train K (2000) Mixed MNL models for discrete response. J Appl Econom 15:447–470 Mesa P, Martín-Ortega J, Berbel J (2008). Análisis multi-criterio de preferencias sociales en gestión hídrica bajo la Directiva Marco del Agua. Revista de Economía Agraria y de los Recursos Naturales 8(2):105–126 Mesa-Jurado AM, Berbel J, Orgaz F (2010) Estimating marginal value of water for irrigated olive grove with the production function method. Span J Agric Res 8(S2) Mitchell R, Carson N (1989) Using surveys to value public goods: the contingent valuation methods. Resources for the Future, Washington, DC Morrison MD, Bennett J, Blamey R, Louviere J (2002) Choice modeling and tests of benefit transfer. Am J Agr Econ 84(1):161–170 Raje DV, Dhobe PS, Deshpande AW (2002) Consumer’s willingness to pay more for municipal supply water: a case study. Ecol Econ 42:391–400 Rolfe J, Bennett J (2009) The impact of offering two versus three alternatives in choice modelling experiments. Ecol Econ 68(4):1140–1148 Rolfe J, Bennett J, Louviere J (2000) Choice modelling and its potential application to tropical rainforest preservation. Ecol Econ 35(2):289–302 Rose JM, Bliemer MCJ (2008) Stated preference experimental design strategies. In: Hensher DA, Button KJ (eds) Handbook of transport modelling. Pergamon Scarpa R, Rose J (2008) Design efficiency for non-market valuation with choice modelling: how to measure it, what to report and why. Aust J Agr Resour Ec 52(3):253–282 Spanish Association for Water Supply and Water Treatment AEAS (2008) Available at: www. aeas.es. Cited Sept 2010 Thurstone LL (1929) Theory of attitude measurement. Psychol Rev 36:222–241 Train K (2003) Discrete choice methods with simulation. Cambridge University Press, Cambridge


1633

Vásques WF, Moxumbder P, Hérandez-Arce J, Berrerns RP (2009) Willingness to pay for safe drinking water: evidence from Parral, Mexico. J Environ Manage 90:3391–3400 Whitehead J, Groothuis PA, Blomquist GC (1993) Testing for non-response and sample selection bias in contingent valuation: analysis of a combination phone/mail survey. Econ Lett 41(2):215– 220 Xenarios S, Bithas K (2009) Welfare improvisation from the receiving waters of urban wastewater systems in the context of the Water Framework Directive. Water Resour Manage (23):981–1000