Assessing the sustainability of water companies: A synthetic indicator approach

Ecological Indicators 61 (2016) 577–587

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Assessing the sustainability of water companies: A synthetic indicator approach María Molinos-Senante a,b,c,∗ , Rui Cunha Marques d , Fátima Pérez e , Trinidad Gómez e , Ramón Sala-Garrido f , Rafael Caballero e a

Departamento de Ingeniería Hidráulica y Ambiental, Pontificia Universidad Católica de Chile, Av. Vicu˜ na Mackenna 4860, Santiago, Chile Escuela de Arquitectura e Instituto de Estudios Urbanos, Pontificia Universidad Católica de Chile, El Comendador 1916, Santiago, Chile c Centro de Desarrollo Urbano Sustentable CONICYT/FONDAP/15110020, Av. Vicu˜ na Mackenna 4860, Santiago, Chile d Center for Urban and Regional Systems (CESUR), CERIS, IST, University of Lisbon, Av. Rovisco Pais, Lisbon, Portugal e Departamento de Economía Aplicada (Matemáticas), Universidad de Málaga, Campus El Ejido, Málaga 29071, Spain f Departamento de Matemáticas para la Economía y la Empresa, Universidad de Valencia, Campus dels Tarongers, Valencia 46022, Spain b

a r t i c l e

i n f o

Article history: Received 25 May 2015 Received in revised form 23 September 2015 Accepted 3 October 2015 Available online 17 November 2015 Keywords: Decision making Multi-criteria decision analysis Sustainability Composite indicator Water companies

a b s t r a c t Performance indicators (PIs) are essential in the benchmarking process used to rate and rank water companies. However, a set of individual PIs does not provide a holistic assessment of company performance from multiple perspectives. A multidimensional evaluation of the performance of water companies can be achieved by aggregating the PIs into a synthetic indicator. Although the concept of sustainability involves economic, environmental and social criteria, most of the previous studies have not considered these three dimensions simultaneously. This paper discusses a process of indicator aggregation using two approaches based on multi-criteria decision analysis to evaluate and compare the sustainability of water companies from a holistic perspective. A synthetic indicator embracing economic, environmental and social PIs was computed for a sample of 154 Portuguese water companies. Both methods yielded similar rankings of water company sustainability. The techniques and results presented in this paper may be utilized as a means of improving the benchmarking process in regulated water industries, as well as providing valuable contributions to decision-makers on the most efficient steps for improving the sustainability of urban water services. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction The benchmarking process in the water industry is currently a topic of international importance. In many countries, such as England and Wales, the Netherlands, Portugal and Chile, the water industry exists as a monopoly or within very restricted power centers. In this context, benchmarking assumes a strategic importance for governments and regulators to create incentives for efficiency and innovation. Without these centralized controls, there is a greater risk that water supply operators would take significant advantage of users by abusing their power within the market place (Marques et al., 2011; Molinos-Senante et al., 2015). Hence,

∗ Corresponding author at: Departamento de Ingeniería Hidráulica y Ambiental, ˜ Mackenna 4860, Santiago, Chile. Pontificia Universidad Católica de Chile, Av. Vicuna Tel.: +56 223544219. E-mail addresses: [email protected] (M. Molinos-Senante), [email protected] (R.C. Marques), f [email protected] (F. Pérez), [email protected] (T. Gómez), [email protected] (R. Sala-Garrido), [email protected] (R. Caballero). http://dx.doi.org/10.1016/j.ecolind.2015.10.009 1470-160X/© 2015 Elsevier Ltd. All rights reserved.

for regulatory purposes, benchmarking is essential to control and supervise the quality of service and/or to establish appropriate tariffs and fair market prices (Marques, 2006). The use of benchmarking is normally based on performance indicators (PIs) which allow for the development of competition by comparison (yardstick competition). One approach to yardstick competition is sunshine regulation which involves a public display of performance data of regulated firms (water companies in our case study), thereby fostering information transparency and allowing consumers/users to make unencumbered comparisons of suppliers. This system encourages the poor-performing water utilities to improve the quality of service they provide since various stakeholders are likely to apply pressure to improve below-average performance (Marques et al., 2011). In the context of regulated water industry, water utilities must provide regular reports on several PIs to the regulator and/or government. These PIs capture management, environmental, financial and, more recently, social data with respect to water operations (Palme and Tillman, 2008). The PI system consists of numerous elements, making them difficult to use by citizens and water

578

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587

regulators for sunshine regulation purposes. One of the difficulties in interpreting PIs is that they are not all the same in terms of importance and this inequality leads to misunderstanding as to the trade-offs among them. That is, high performance on one indicator does not necessarily compensate for low performance on another. A major limitation of evaluating performance based on a set of PIs, therefore, is that it does not provide a holistic view. Using this approach, the output of the assessment process is not a measure of general performance, and it is therefore very difficult to rank water companies based on their general performance (Duarte et al., 2009). One way of bypassing this limitation is to aggregate the PIs, converting them into a synthetic indicator which provides a multidimensional assessment of the performance of water companies. Aggregating individual indices using Multi-Criteria Decision Analysis (MCDA) is a common technique to construct synthetic indices for sustainability evaluation. MCDA has been used to evaluate the sustainability of a wide range of activities, services and/or processes (e.g., Giannetti et al., 2009; Voces et al., 2012; Molinos-Senante et al., 2014). At the same time, consideration of sustainable development has become an increasingly important factor for the regulation and governance of the water industry (Cashman and Lewis, 2007). While there is an increasing recognition of the need to improve the sustainability of urban water systems, previous studies suggest a lack of consensus on the appropriate criteria to assess water companies’ performance with regard to sustainability (e.g., Foxon et al., 2002; Sahely et al., 2005; Palme and Tillman, 2009; Rojas-Torres et al., 2014). This is due, in part, to the lack of a clear definition of a sustainable water company. Nevertheless, for this study we adopted the traditional vision of sustainability which separates the concept of sustainability into three dimensions: environmental, economic and social (WCED, 1987; Singh et al., 2012). The majority of published studies on urban water supply have focused on assessing the sustainability of physical and engineering aspects of water supply systems, in particular the water distribution networks (e.g. Hamouda et al., 2009; Tabesh and Saber, 2012; Marques et al., 2015; Aydin et al., 2014), leading to sustainability indices based primarily on technical criteria such as reliability, resiliency and vulnerability. Studies have also evaluated sustainability from an environmental perspective using life cycle assessment (Lundie et al., 2010; Schulz et al., 2012). The economic sustainability of water distribution systems have been evaluated using both a life cycle costing approach (Schulz et al., 2012) or minimizing overall system costs (Ahn and Kang, 2014). There is, however, a lack of information in the published literature that focuses on the sustainability of water companies themselves. Based on our review, only four papers (Klostermann and Cramer, 2007; Duarte et al., 2009; Singh et al., 2010; Marques et al., 2015) describe the application of empirical data to evaluate the sustainability of a sample of water companies; only two of them used MCDA (Duarte et al., 2009; Marques et al., 2015). While Klostermann and Cramer (2007) compared the sustainability of two Dutch water companies using several PIs, the lack of an integrated evaluation process relative to the selected companies makes the PIs difficult to utilize on a broader scale by key decision makers. Singh et al. (2010) evaluated the sustainability of 18 Indian water companies using six sustainability parameters. To accomplish this, they aggregated technical efficiency and scaled efficiency scores computed by data envelopment analysis methodology. Duarte et al. (2009) proposed a synthetic index of service quality to evaluate the general performance of water companies, where the index is calculated as a weighted linear combination of the normalized scores of each performance indicator. The initial indicators are grouped so as to reflect protection of the water users´ı interest, as well as sustainability of the utility and environment. These factors do not correspond to the traditional dimensions of sustainability (economic, social

and environmental). From a methodological point of view, the normalization process was based on “fuzzy” data sets and the opinion of a panel of experts by using the analytic hierarchy process (AHP). Although the AHP has several advantages, there are downsides as well. First, the number of required pairwise comparisons may become very large, making the AHP a lengthy and potentially cumbersome task. Second, distinguishing among preferences using Saaty’s scale may prove difficult for the decision maker. Saaty’s scale consists of 1–9 ratios, each indicating how many times one element is more important or dominant over another element with respect to the criterion to which they are compared (Saaty, 2008). Third, since the results of the AHP are dependent on the participants who perform the pairwise comparisons, those results will always be subject to human error and a certain level of subjectivity that varies from person to person (Molinos-Senante et al., 2014). Finally, the study by Marques et al. (2015) evaluated the sustainability of water supply systems through the Macbeth method, adopting the water utility of Lisbon, Portugal as a case study. Therefore, in order to weight the coefficients as required by the AHP, stakeholders are asked to pass judgment on the difference in attractiveness between two criteria at a time using a semantic scale having seven categories. The performance measures are usually qualitative judgements which are further quantified proportionally on a 0–100 scale (Bana e Costa and Vansnick, 1994). In addition to employing the MCDA methodology, these researchers adopted two additional dimensions of sustainability, consisting of assets and governance, which were employed along with the previously indicated metrics of environment, social and economic dimensions. The goal of the study described here is to contribute to an improvement in the sustainability assessment system of water companies. In doing so, an initial set of sustainability indicators is aggregated into a synthetic indicator based on two novel methodologies, specifically the distance-principal component (DPC) and the global programming synthetic indicator (GPSI), yielding a general measure of sustainability for each water company. Since DPC follows a statistical approach while GPSI in based on a non-statistical technique, the final metric provides insight into differences in ranking of water companies when the two different approaches are applied in concert. Information provided in this paper provides a means for the assessment of the reliability of the DPC and NGPSI approaches and contributes to the standardization of the two methods. Although there have been a significant number of empirical studies recently completed that assess water company performance, none of them focus on developing a composite indicator which provides a holistic performance perspective. The current study, therefore, is a pioneering and novel approach in the framework of water company performance through its integration of multiple indicators into a single and synthetic metric. This synthetic indicator facilitates interpretation of the PIs and allows water companies to be ranked by sustainability, thus improving benchmarking and allowing both societal users and water managers to make critical decisions based on quantitative data and, if needed, implement corrective measures to improve the sustainability of water urban services over time. Following this Introduction, the paper is divided into four additional sections. Section 2 describes the methodologies applied. Sections 3 and 4 illustrate the case study and then provide results and discussion, respectively. Finally, conclusions are presented in Section 5.

2. Methodology Previous studies have illustrated the multiple methodological approaches for constructing synthetic or composite indicators

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587

(Nardo et al., 2005; Blancas et al., 2010a). A method is selected by the researcher based on the study goals and characteristics of the data and analysis. As there is no consensual theoretical framework for obtaining synthetic measures, we applied a statistical approach (DPC) and a non-statistical technique (GPSI) to evaluate the sustainability for a set of water companies. The computation of both indicators allowed us to compare and contrast the results under both approaches (statistical and non-statistical). The main advantages of constructing the synthetic indicator using DPC methodology are (Blancas et al., 2010a, 2011; Caballero Fernandez and Cruz Morato, 2011): (i) the values of the indicators are easy to interpret because they are always positive, which facilitates the comparison among the units evaluated (water companies in our case study); (ii) the weights for each initial indicator are determined endogenously. Thus, additional information is not required by the analyst or by a panel of experts, resulting in a less subjective composite indicator; (iii) because of the weighting system, more influential indicators affecting sustainability are more easily identified. The three main advantages of the GPSI approach over the statistical methods are: (i) GPSI does not require the initial system of indicators to be normalized; (ii) GPSI can be constructed even when the number of units to be evaluated is lower than the number of initial indicators; (iii) the composite indicator is constructed using all the initial indicators without losing information. However, the limitation of this approach is that it requires setting weights and aspiration levels for each initial indicator, which introduces some subjectivity into the analysis (Blancas et al., 2010b).

2.1. Distance-principal component (DPC) The DPC indicator was proposed by Blancas et al. (2010a) to assess the sustainability of tourism destinations. It is characterized by combining two techniques, such as principal component analysis (PCA) and distance to a reference point. In this statistical method the weight of each indicator is calculated based on the variability of the data. PCA is widely used to construct synthetic indicators across a variety of applications (Somarriba and Pena, 2009; Jiang et al., 2013; Reisi et al., 2014; Salvati and Carlucci, 2015). It can be used as a means of representing the majority of information available for a system through a limited number of variables which are linear combinations of the original variables. The main advantage of this method is that it is not necessary to apply a weight to each component because that aspect is already inherent to the PCA. That is, each initial indicator is weighted according to the quantity of the information system explained by each variable and the contribution of each initial indicator to the variance of the data, thereby reducing the subjectivity of the aggregation process (Pérez et al., 2013; Blancas et al., 2011). Nevertheless, interpreting PCA results for each observation can often be difficult since the main components are linear combinations of the initial indicators. To address this issue, Blancas et al. (2010a) combined PCA with the distance to a reference point, based on a multi-criteria decision-making approach. The procedure to obtain a synthetic indicator based on DPC methodology can be summarized in five main steps. For illustration, consider a system composed of m indicators to assess the synthetic sustainability of n water companies. Let Iij denotes the value of the ith water company in the jth indicator. The first step in obtaining the synthetic indicator is to distinguish between positive and negative indicators based on the direction of change. Positive indicators are those for which a larger value reflects an improvement in the sustainability of the water company, e.g., an ability to internally generate energy for its own use. On the other hand, negative indicators are those for which larger values reflect a decline in the

579

sustainability of the water company, e.g., leakage(Molinos-Senante et al., 2014). The second step is to homogenize the direction of improvement for all initial indicators. The goal is to prevent the positive and negative indicators from balancing (net zero change) since, as it is shown in Eq. (2), the synthetic indicator is the weighted addition of the initial indicators. Following Blancas et al. (2011), negative indicators are converted into positive indicators by changing the value sign. Third, the values of the initial indicators are defined as the distance of each water company relative to a fixed reference point. Bearing in mind that after the second step all initial indicators are positive, the minimum value of each indicator is taken as a reference value. Hence, by measuring the distance to the minimum value, the distance to the lower bound is obtained. A larger distance indicates higher sustainability. Fourth, the data are normalized so that the units used to measure the indicator do not affect the results. To do this, the distance to the lower bound is divided by the difference between the maximum and the minimum value as shown in Eq. (1): Iij =

Iij − Ijmin Ijmax

(1)

− Ijmin

where INij is the normalized value of the ith water company in the jth indicator and Ijmin and Ijmax are the minimum and maximum values, respectively, of the jth indicator. The fifth and final step is to apply the PCA to define the weights assigned to each initial indicator endogenously. According to Blancas et al. (2010a), the weight of each indicator is calculated as the product of the variance explained by each selected principal component and the absolute value of the correlation of each indicator with each adopted principal component. Therefore, the weight applied to each initial indicator is based on the quantity of the information system explained by each component and the contribution of each initial indicator to this variance. The composite indicator is formulated as in equation (2) (Blancas et al., 2011): DPCi =

m  j=1



INij

 q 

  VEk Corrkj 



for i = 1, 2, . . ., n

(2)

k=1

where n is the number of water companies; m is the number of initial indicators; q is the number of principal components chosen; VEk is the explained variance for component kth; and Corrkj is the correlation of initial indicator jth with the principal component kth. 2.2. Goal programming synthetic indicator (GPSI) The second approach used in the development of a synthetic indicator is based on goal programming. Whereas the DPC approach is a statistical one, GPSI is a non-statistical technique in which synthetic measures are obtained by deviation variables associated with the goals defined for each initial indicator (Diaz-Balteiro and Romero, 2004; Lozano-Oyola et al., 2012). In this context, we followed the methodological approach proposed by Blancas et al. (2010b). This method enables us to construct two different indicators depending on the degree of compensation allowed between the initial indicators, such as the restrictive goal programming synthetic indicator (RGPSI) and the net goal programming synthetic indicator (NGPSI). RGPSI does not allow for compensation between strengths and weaknesses and, therefore, provides for a final ranking that reflects strong sustainability. In contrast, NGPSI assesses the sustainability of each unit by aggregating its strengths and weaknesses and therefore is considered a measure of weak sustainability. In order to improve the comparability of the two approaches (statistical and non-statistical), information was aggregated by using NGPSI.

580

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587

To construct the NGPSI from the DPC indicator (using the parameters outlined earlier), the analyst must distinguish not only between positive and negative indicators but also neutral ones. An indicator is neutral when a specific value is desired and, therefore, positive and negative deviations are undesirable (e.g., pH of the water). Three types of indicators are possible: Iij+ which represents

− the ith water company in the jth positive indicator; Iik which provides the kth negative indicator for ith water company; Iih which provides the hth neutral indicator for the ith water company. The second step was to define an aspiration level for each indicator, i.e., u+ , u− and uh for positive, negative and neutral indicators, j k respectively. Aspiration levels are interpreted as follows: (i) for positive indicators, u+ is the minimum level at which a water comj pany exhibits a suitable solution regarding the issue evaluated by is the maximum the indicator jth; (ii) for negative indicators, u− k level at which a water company presents a favorable situation with respect to the aspect assessed by the indicator kth; (iii) for neutral indicators, uh is the specific value desired for the indicator hth. In other words, unlike reference points, aspiration levels are defined differently for positive, negative and neutral indicators. Thus, the aspiration levels are the thresholds that are considered when determining if a unit (water company in our case study) is sustainable. Two main approaches can be used to determine aspiration levels: (i) use of a panel of experts or (ii) the mean of each indicator. The first approach is more suitable when the stakeholders are interested in directly setting the thresholds for sustainability. Since in our study this information was unavailable, aspiration levels were fixed following the latter approach, i.e., based on the mean values of each indicator. It is important to note that the set value for the aspiration level depends on the requirements that the analyst imposes on a water company to consider it “sustainable.” Blancas et al. (2010b) and Lozano-Oyola et al. (2012) suggested 80% of the mean as an adequate aspiration level. In the case of negative indicators, water companies should strive to reduce the indicator value (i.e., improving their sustainability) and, therefore, the value of the aspiration levels should be greater than 80%. To be consistent with the positive indicators, the aspiration levels for negative indicators are defined as the reciprocal of 80% of the mean, i.e., 1/0.8 which is 125% of the mean. Third, since NGPSI is based on goal programming, the initial indicator value for each water company was compared to its aspiration level. For each indicator a goal is defined using deviation variables denoted by n and p. These variables indicate the difference between the value of an indicator and the corresponding aspiration level for each water company. For the ith water company, the goals are defined as (Blancas et al., 2010b):

For positive indicators :

Iij+ + n+ − p+ = u+ with n+ , p+ ≥ 0; ij ij j ij ij

n+ · p+ =0 ij ij

For negative indicators :

(3)

− Iik + n− − p− = u− with n− , p− ≥ 0; ik ik k ik ik

n− · p− =0 ik ik

For neutral indicators :

(4)

Iih + nih − pih = uh with nih , pih ≥ 0;

nih · pih = 0 where

n+ ij

is the negative deviation variable and

certain deviation variables are determined to be null will depend on the improvement direction of each indicator. In the case of positive indicators, negative deviation variable (n+ ) is non-desirable since ij the better-positioned water companies with more progressive programs will achieve the aspiration level or higher. For negative indicators, the non-desirable variable will be the positive deviation variable (p− ), allowing better-positioned water companies to ik achieve the aspiration level or lower. Finally, in the case of neutral indicators, both deviation variables (nih and pih ) are non-desirable. The fourth step was to apply an appropriate weight to the relative importance of each initial indicator. Therefore, wj refers to the weight assigned to the jth indicator. Finally, the NGPSI was defined by using degree of the deviation between the values of each indicator with their aspiration levels. The NGPSI provides information about the relative position of each water company without t time-consuming task of completing all of the aspiration level input. The NGPSI can be divided into two components, namely NGPSI+ and NGPSI− . The first (NGPSI+ ) refers to the strengths of each water company in relation to the aspect assessed. Its definition is based on the aggregation of deviation variables where a higher value indicates a better relative position, i.e., for positive indicators and n− for negative indicators. NGPSI+ is p+ ij ik calculated as follows: NGPSIi+ =

 wj p+ij j∈J

u+ j

+

 wk n− k∈K

u− k

ik

∀i ∈





1, 2, . . ., n

(6)

The J is the set of positive initial indicators and K is the set of negative initial indicators. The second NGPSI component (NGPSI− ) measures the weaknesses of each water company relating to the indicator system. It provides information about the degree to which a water company does not fulfill the set of aspiration levels. Its formulation is similar to NGPSI+ , but it accounts for the fact that neutral indicators are represented only by their weaknesses since their deviation variable only indicate weakness. NGPSI− is calculated as follows: NGPSIi− =

 wj n+ij j∈J

∀i ∈



u+ j

+

 wk p− k∈K

u− k



1, 2, . . ., n .

ik

+

 w (n + p ) h ih ih h∈H

uh

(7)

where H is the set of neutral initial indicators and the remaining variables are defined as described earlier. Based on the NGPSI+ and the NGPSI− , the NGPSI for the water company i is determined as follows: NGPSIi = ˛NGPSIi+ − ˇNGPSIi−

(8)

where ˛ and ˇ are relative weights assigned to the strengths and the weaknesses of the water companies. Although from a theoretical point of view, ˛ and ˇ can be different from one, following Blancas et al. (2010) we considered the strengths and the weaknesses of each water company to be of equal importance and, therefore, ˛ = ˇ = 1.

(5) p+ ij

3. Case study is the positive

deviation variable associated with the positive indicator; n− is the ik negative deviation variable and p− is the positive deviation variable ik associated with the negative indicator; nih is the negative deviation variable and pih is the positive deviation variable associated with the neutral indicator. As calculated using Eqs. (3)–(5), whether

We evaluated the sustainability of 154 Portuguese water companies with regard only to the water services they provide. Information about the initial set of indicators is available at ERSAR (Portuguese national regulator) webpage for 12 which is available from its webpage.

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587

581

Table 1 Models of management in the Portuguese water sector.

Retail Only water Water and wastewater Water and others Wholesale Only water Water and wastewater

Total (no.)

Concessionaire (no.); (%)

Mixed company (no.); (%)

Municipal company (no.); (%)

PuPs (no.); (%)

Semiautonomous (no.); (%)

Direct provision (no.); (%)

265 11 43 211 14 4 10

27; 10.1 6; 54.5 21; 48.8 – 1; 7.1 1; 25.0 –

9; 3.3 1; 9.1 1; 2.3 7; 3.3 – – –

15; 5.6 – 8; 18.6 7; 3.3 1; 7.1 1; 25.0 –

2; 1.0 – 2; 4.7 – 12; 85.8 2; 50.0 10; 100.0

21; 7.9 4; 36.4 11; 25.6 6; 2.8 – – –

191; 72.1 – – 191; 90.6 – – –

3.1. Description of the water industry in Portugal Approximately 10.5 million people live in Portugal where three major issues mark the water industry. The first one is the separation of the wholesale and retail markets, both in water and wastewater. In Portugal there are distinct companies that provide water and/or wastewater services. Some of them provide services in the wholesale segment (all activities involved in the urban water cycle) while others focus on the retail segment (distribution of drinking water and collection of wastewater). The wholesale companies, in general, are organized through public partnerships between the local governments (municipalities) and the central Government. The retail companies are usually the responsibility of the municipalities (Pinto et al., 2015). The second issue concerns the diversity of existing unbundling models. In addition to the public partnerships between municipalities and the central Government, the municipalities can engage directly with the private sector in concession arrangements (Marques and Berg, 2011), create municipal companies which can be wholly municipal or involve the private sector in a mixed arrangement (Da Cruz and Marques, 2012), and provide the water-related service (to users) together with other municipal activities, or create a semi-autonomous but non-corporate structure that has some financial and administrative independence. In the retail segment the municipalities can also enter into partnerships with the state (called PuPs). This water industry structure is quite distinctive in Portugal with only one other European country, Ireland, having similar management models (De Witte and Marques, 2012). Finally, the third distinctive feature is the existence of a national regulatory authority. This is atypical of water sector groups in Europe, except for the regulatory structure in England and Wales. Other European countries such as the Netherlands and Italy do have a regulatory system, but it is not so strict and standardized as it is in Portugal, England or Wales. Moreover, other European countries such as Denmark, France and Germany have a large degree of regulatory decentralization and industry fragmentation (Boscheck et al., 2013). The regulatory model in Portugal has some outstanding positive points, in particular the quality of service and technical regulation, as well as access to information. There are, unfortunately, several drawbacks including poor governance and failure to address identified problems. Its structure and functionality has limited and constrained private sector participation which could lead to more efficient models and development of the sector. The quality of service regulation is based on the sunshine regulatory method which has had very positive results (Marques and Simões, 2008; Simões and Marques, 2012). Table 1 lists the primary aspects of the Portuguese water sector. The evolution of the Portuguese water industry has been widely studied, not only because most of the performance indicators are public, but also because it is one of the first countries in the world to have implemented sunshine regulation. Some previous

studies (Marques, 2006; Marques and Simões, 2008; Simões and Marques, 2012; Carvalho and Marques, 2015) reported that the water sector has made noticeable progress in the past two decades by increasing the coverage and quality of the services provided. Nevertheless, some problems persist, such as the high and moderate water losses, poor staff productivity, weaknesses in wastewater treatment and lack of cost recovery with significant dependence on external funds. The sector is also know for significant politicization of the sector, both at the Central State and local government levels. 3.2. Selection of sustainability indicators The selection of the set of initial indicators for use in constructing a synthetic indicator is an essential step (Balkema et al., 2002) since the general sustainability of a water company will depend on these indicators. There is no a single set of valid indicators to assess sustainability, although EPA (2012) proposed some indicators to assess water utility sustainability grouped in the three traditional dimensions (environmental, social and economic) used to characterize the concept of sustainability. Previous studies evaluating the sustainability of processes and/or organizations, indicated the selection of the initial indicators should be based on the following criteria: representativeness, relevance, reliability, sensitivity, ease of understanding, comparability and transparency (Perez et al., 2015; Blanc et al., 2008). The indicators, therefore, should be capable of expressing accurate progress toward or away from sustainability in a manner that can be easily interpreted (Molinos-Senante et al., 2014). Moreover, the selection of initial indicators is strongly related to the availability of statistical data (Blancas et al., 2011). In our case study the initial system of indicators was designed as a balance between the relevance of the indicators and their availability. For each dimension of sustainability (economic, environmental and social), several indicators were identified. We defined our system by reviewing data and selecting from the ERSAR list of Portuguese water companies for 2012. In particular, our initial indicator system was composed of 14 sustainability indicators which are presented in Table 2, along with a short summary of each indicator detailing its acronym, direction of improvement (positive, negative or neutral indicators), associated evaluation issues on which the system is built and units of measure. The assignment of sustainability indicator to each dimension was conducted by considering its characteristics. Table 3 shows the main statistical features of each initial indicator. A brief description of the sustainability indicators for the three dimensions is given below: Social dimension: Five indicators were selected that should allow managers and water regulators to evaluate the social benefits of the water supply service. The following social indicators were considered:

582

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587

Table 2 Set of indicators to assess the sustainability of water supply companies. Dimension

Indicator

Acronym

Direction

Formula

Unit

Social

Service coverage Safe drinking water Capacity of reserve Certification of occupational risks and health Other certifications

ISS1 ISS2 ISS3 ISS4

Positive Positive Positive Neutral

ISS1 = (Number of household supplied)/(Total number of households) × 100 ISS2 = (Satisfactory analysis for quality parameters)/(Total number of analysis) × 100 – –

% % Days –

ISS5

Neutral

–

Environmental

Water losses Production of energy Use of energy for water pumping Certification of environmental issues Certification of water quality issues

IEN1 IEN2

Negative Positive

IEN3

Negative

IEN2 = (Production of energy by the company (kWh/year))/(Total consumption of energy (kWh/year)) × 100. IEN3 = (Energy consumed (kWh/year))/(Normalization factor (m3 /year)) × 100m.

IEN4

Neutral

–

–

IEN5

Neutral

–

–

Non-revenue water Staff ratio

IEC1 IEC2

Negative Negative

IEC1 = (Total water suppplied (m3 /year))/(Total water revenued (m3 /year)) × 100. IEC2 = (Number of employees)/(103 number of connections).

Operating cost coverage ratio Index of knowledge

IEC3

Positive

IEC4 = (Total annual operational revenues)/(Total annual operational costs) × 100

% Number/103 connections %

IEC4

Positive

–

–

Economic

m km day % kWh/(m3 100 m)

More information about the definition of each indicator is available at the web page of ERSAR: http://www.ersar.pt/website/.

Table 3 Main descriptive statistics. Indicator

Average

Service coverage (%) Safe drinking water (%) Capacity of reserve (days) Certification of occupational risks and health* Other certifications * Water losses (m3 /km day) Production of energy (%) Use of energy for water pumping (kWh/(m3 100 m) Certification of environmental issues* Certification of water quality issues* Non-revenue water (%) Staff ratio (Number/103 connections) Operating cost coverage ratio (%) Index of knowledge

92.9 98.6 1.6 7.8 8.4 126.9 6.1 0.9 9.7 25.3 36.7 2.4 95.6 48.8

SD

Minimum

Maximum

9.8 1.9 1.1

54.0 87.6 0.0

100.0 100.0 6.3

82.1 68.7 1.2

0.4 0.0 0.3

386.0 653.3 12.8

13.1 1.3 41.5 24.2

6.8 0.8 10.0 10.3

74.6 10.1 320.0 100.0

Source: ERSAR (Portuguese national regulator) * The percentage of water companies which have certification.

(ISS1 ) Service coverage (%): The percentage of the total number of households located in the intervention area for which the water supply company provides effective connected service. (ISS2 ) Safe drinking water (%): The percentage of water supplied that meets the legal requirements regarding quality parameters. (ISS3 ) Reserve capacity of treated water (days): The capacity of the water company to supply water to customers if new water resources are not available. (ISS4 ) Certification of management systems for occupational risk and health issues at work. (ISS5 ) Other certifications involving multiple issues such as corporate social responsibility and consumer protection mechanisms. Environmental indicators: Are related to environmental impacts associated with water treatment and supply, as well as the management of the owned resource. The following environmental indicators were considered: (IEN1 ) Water losses in the network (m3 /km·day): Leakages are negative environmental impacts related not only to loss of water

but also with loss of treatment reagents and energy (HernándezSancho et al., 2012). (IEN2 ) Internal power generation by the water company (%): Water companies that generate a greater percentage of energy (relative to their own needs) are not only less dependent on external power sources but are also more sustainable from an environmental perspective. (IEN3 ) Energy efficiency for water pumping (kWh/(m3 ·100 m): This measures the average consumption of energy for water pumping, taking into account the depth of the well. (IEN4 ) Certification of management systems for issues such as environmental responsibility and environmental impact assessment mechanisms. (IEN5 ) Certification of management systems for water quality issues. Economic indicators: The asset management costs within the water company and the operational costs of providing water supply services. The specific economic indicators that were selected are as follows:

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587

(IEC1 ) Non-revenue water (NRW) (%): Defined as the percentage of water that it is supplied but it is not invoiced. (IEC2 ) Adequacy of staffing (number/1000 connections): Defined as the number of full-time equivalent (FTE) employees per 1000 connections. This is a proxy of staff costs. (IEC3 ) Operating cost coverage ratio: The ratio between total annual operational revenues and total annual operational costs. (IEC4 ) Index of knowledge about infrastructure and assets management: An index ranging between 0 and 100, calculated by ERSAR based on the information available about infrastructure, interventions performed and level of asset management conducted by the water company. The indicators ISS4 , ISS5 , IEN4 and IEN5 are binary and have a value of either 1, if the water company has the certification, or 0, if no certification is present. 4. Results and discussion 4.1. Application of the DPC To conduct the empirical analysis, we first confirmed the suitability of the database for PCA. In doing so, we applied the Kaiser–Meyer–Olkin (KMO) Measure of Sampling Adequacy and the Barlett Test of Shericity (Wong and Pang, 2003). KMO is a summary of how small the partial correlations are relative to the original (zero-order) correlations. If the variables share common factor(s), then the partial correlations should be small and the KMO should be close to 1.0. KMO values greater than 0.8 can be considered favorable. The Barlett Test of Shericity checks the hypothesis that the correlation matrix is an identity matrix implying that all of the variables are uncorrelated. If the significance value for this test is less than the selected alpha level, the null hypothesis that the population matrix is an identity matrix can be rejected and it is concluded that there are correlations in the data set that are appropriate for factor analysis. Results from both methods confirmed that our database was suitable to perform a PCA. The weights attributed to each initial indicator within each dimension provide some insight into the factors that contribute most to sustainability (Table 4). For six out of the 14 initial indicators, the weights are lower than 0.25. These six indicators are divided evenly among Environmental (self-production of energy and use of energy for water pumping), Economic (operating cost coverage ratio and staff ratio) and Social (safe drinking water and reserve capacity). In contrast, the indicators that were assigned the highest weights are those associated with certification issues. In particular, the certification of water quality and environmental

Table 4 Weights of the sustainability indicators by DPC approach. Dimension

Indicator

Weight

Social

Service coverage (ISS1 ) Safe drinking water (ISS2 ) Capacity of reserve (ISS3 ) Certification of occupational risks and health (ISS4 ) Other certifications (ISS5 )

0.3306 0.2160 0.2144 0.3696 0.3405

Environmental

Water losses (IEN1 ) Production of energy (IEN2 ) Use of energy for water pumping (IEN3 ) Certification of environmental issues (IEN4 ) Certification of water quality issues (IEN5 )

0.2893 0.1969 0.2294 0.3809 0.4173

Economic

Non-revenue water (IEC1 ) Staff ratio (IEC2 ) Operating cost coverage ratio (IEC3 ) Index of knowledge (IEC4 )

0.3356 0.2081 0.2007 0.3306

583

Table 5 Synthetic index of the sustainability based on DPC approach for the best and worst 20 water companies from Portugal. Water company

Synthetic indicator

Ranking

Águas de Cascais Águas de Valongo Águas de Paredes Águas e Parque Biológico de Gaia SMSB de Viana do Castelo Indaqua Matosinhos EPAL Águas de Alenquer AGERE INOVA Aquamaior Águas de Mafra VIMÁGUA Aquaelvas INFRAMOURA INFRAQUINTA Tavira Verde Águas da Figueira CM de Pombal SMAS de Sintra CM de Portel CM de Ponte da Barca CM de Reguengos de Monsaraz CM de Arronches CM de Seia CM de Pinhel CM de Paredes de Coura CM de Vila Nova de Famalicão CM de Lamego CM de Celorico da Beira CM de Avis CM de Alijó CM de Murc¸a CM de Castanheira de Pera CM de Vila Nova de Poiares CM de Nelas CM de Mourão CM de São João da Pesqueira CM de Baião EPMAR Vieira do Minho

3.1752 3.0145 2.8178 2.7856 2.7395 2.7385 2.7329 2.7091 2.6233 2.6198 2.5353 2.5308 2.4912 2.3267 2.3057 2.2674 2.1503 2.1319 2.0754 2.0686 1.1078 1.1043 1.0987 1.0977 1.0977 1.0787 1.0684 1.0677 1.0569 1.0384 1.0323 1.0091 0.9656 0.9538 0.9404 0.9123 0.8671 0.8322 0.8159 0.7778

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154

issues are the two initial indicators contributing the most to water company sustainability. From a policy perspective, these results illustrate the importance of acknowledging good performance of water companies through the use of certification programs. Such programs not only provide visual and physical evidence of performance, but also will force companies to make necessary improvements in order to obtain the certification. Regulators should promote those programs with proven economic incentives and/or benefits for certified water companies during the benchmarking process. Given that the empirical analysis conducted here encompasses a large number of water companies, the values obtained for the general sustainability for all companies are presented as supplementary information. In order to compare the results when general sustainability is computed by using the goal programming approach (NGPSI), the results of the 20 best and worst water companies are presented. Based on the DPC approach, the most sustainable water company is Águas de Cascais with a synthetic index of sustainability of 3.18 (Table 5). By contrast, the least sustainable company is EPMAR Vieira do Minho with a sustainability index of 0.78. There are significant differences in the sustainability of Portuguese water companies (Fig. 1), and substantial room for improvement. When the synthetic sustainability index is normalized (value of 1 assigned to the most sustainable water company and 0 assigned to the least sustainable company), only a small number of water companies can be clearly classified as highly sustainable (Fig. 2).

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587 3.5

70

3.0

60 Global index of sustainability

GLobal indicator of sustainability

584

2.5 2.0 1.5 1.0 0.5

Fig. 1. Synthetic index of the sustainability based on DPC approach for the Portuguese water companies evaluated.

60 % of water companies

30 20 10

-10

Water companies

50 40 30 20 10 0 0.0 <= SI < 0.2 0.2 <= SI < 0.4 0.4 <= SI < 0.6 0.6 <= SI < 0.8 0.8 <= SI <= 1.0 Global sustainability index (SI)

Fig. 2. Water companies grouped by synthetic sustainability index based on DPC. 0.7

Global Sustainability Index

40

0

0.0

0.6 0.5 0.4 0.3 0.2 0.1 0.0

50

Concession

Municipal company

Semi-autonomous service

Direct service

Fig. 3. Sustainability of water companies grouped by models of management.

Specifically, only five out of 154 water companies have a normalized sustainability index of 0.8 or higher. A very large proportion (76%) of the water companies had a normalized sustainability index of less than 0.4. These scores emphasize the great potential for improvements in sustainability for the vast majority of water companies. It is interesting to note that most of the more sustainable water utilities (higher indices) are privately owned, which corroborates the results of previous studies in Portugal which highlight the superior performance of privately-held companies (Marques, 2008). In general, the municipal companies also yielded good performance results. In contrast, the direct control of the water supply by the municipal departments provides evidence of a poorly performing water service system which is statistically less sustainable than the concessions and municipal companies (Fig. 3). 4.2. Application of the NGPSI To calculate the general sustainability of water companies using the goal programming approach, i.e., to compute the NGPSI, we first established the aspiration levels for each initial indicator. Due to the limited information available on this study topic, we were unable to use an external source as a reference, and therefore

Water companies

Fig. 4. Synthetic index of the sustainability based on GPSI approach for the Portuguese water companies evaluated.

followed the direction guidance provided by Blancas et al. (2010b) and Lozano-Oyola et al. (2012). Given the large number of water companies evaluated, we selected aspiration levels based on the values of each indicator. Specifically, the aspiration levels of the positive indicators were fixed at 80% of their mean values, and for negative indicators, the reciprocal percentage of the mean values was used. For assigned indicator weights, the same importance was assigned to each indicator. To enhance and improve the public participation in the definition and assessment of the sustainability of water companies, the aspiration levels and the weights of the initial indicators should be assigned with consideration of the opinions of the local population, water managers and regulatory officials. It should be noted that unlike DPC, NGPSI follows a non-statistical approach and, therefore, the data of the sample itself cannot be used to establish indicator weights. Moreover, since one of the objectives of this study was to assess ranking stability when different methods were used, the same weight was assigned to all indicators under the NGPSI approach. With regard to the synthetic indicator of sustainability based on GPSI, the 20 best and worst water companies, based on the NGPSI ranking, are shown in Table 6. This indicator shows that the most sustainable water company is Indaqua Matosinhos while the least sustainable is CM de Mourão. Using these same data, 56 of 154 water companies (36.4%) have a positive value for the NGPIS. This information indicates the strengths of these water companies outweigh their weaknesses. The opposite is true with the remaining water companies (63.6%) which demonstrate a negative value for the NGPSI. In order to test the similarity of the water company rankings obtained through the DPC and NGPSI approaches, Spearman’s rho correlation coefficient was computed. This is a non-parametric measure of statistical dependence used to assess how well the relationship between two variables can be described using a monotonic function. If each of the variables is a perfect monotone function of the other, then the Spearman’s rho correlation coefficient is +1 or −1 (Corder and Foreman, 2014). In our case study, the Spearman’s rho correlation coefficient between DPC and NGPSI values was 0.937, which indicates both approaches provide a very similar ranking of water companies regarding their sustainability. Selection of the best approach to assess sustainability, therefore, will depend on several factors, including: the number of units to be evaluated, the participation of the local population and stakeholders for indicator weight assignment and the a priori goal of the sustainability assessment. Given a thorough understanding of these factors, the resulting data may be used by water companies to improve their sustainability, by regulators during the benchmarking process, by interested parties to identify factors with the greatest contribution to sustainability or to create social awareness around the sustainability issue. In

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587

585

Table 6 Synthetic index of the sustainability based on goal programming approach for the best and worst 20 water companies from Portugal. Water company

NGPSI+

NGPSI−

NGPSI

Ranking

Indaqua Matosinhos Águas de Cascais Tavira Verde Águas e Parque Biológico de Gaia Águas de Alenquer Águas de Valongo VIMÁGUA Águas de Paredes SMSB de Viana do Castelo INOVA AGERE EPAL Águas de Barcelos Águas de Mafra Águas de Gondomar SMAS de Sintra Aquaelvas Aquamaior CM de Miranda do Corvo AGS – Pac¸os de Ferreira CM de Vila Nova de Famalicão CM de Seia CM de São Brás de Alportel CM de Estremoz CM de Ferreira do Alentejo CM de Portel CM de Paredes de Coura CM de Mértola CM de Castro Verde CM de São João da Pesqueira CM de Celorico da Beira CM de Arronches CM de Vila Nova de Poiares CM de Murc¸a CM de Avis EPMAR Vieira do Minho CM de Baião CM de Castanheira de Pera CM de Nelas CM de Mourão

64.64 49.96 44.02 41.64 39.12 37.70 36.54 36.55 34.92 34.19 34.09 34.12 29.04 24.65 21.57 21.40 20.18 20.45 20.00 20.87 0.96 1.33 1.14 1.16 0.92 1.54 1.99 1.96 1.25 1.18 1.08 0.96 2.07 2.78 0.87 0.72 0.73 0.63 0.63 0.80

1.39 0.16 3.00 1.00 1.00 2.00 2.05 2.07 2.10 2.23 2.49 3.35 4.76 3.00 4.00 3.87 3.00 3.37 3.00 4.00 5.83 6.22 6.13 6.18 5.97 6.62 7.11 7.16 6.64 6.68 6.64 6.57 7.78 8.60 6.93 6.87 7.26 7.40 7.59 7.88

63.25 49.80 41.02 40.64 38.12 35.70 34.50 34.48 32.82 31.96 31.59 30.77 24.28 21.65 17.57 17.53 17.18 17.08 17.00 16.87 −4.87 −4.89 −4.98 −5.02 −5.05 −5.08 −5.11 −5.20 −5.38 −5.49 −5.56 −5.62 −5.71 −5.81 −6.06 −6.15 −6.53 −6.78 −6.96 −7.08

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154

summary, to select the appropriate method to construct the composite indicator, the analyst should consider the context of the assessment and the main advantages of the different methodological approaches. All studies have certain limitations and ours is no exception. It should be noted that the synthetic indicator constructed in this paper integrates 14 basic indicators and is limited by the information those indicators provide. Future research of this issue should focus on integrating additional indicators into the assessment, thus providing greater precision about water company performance. Moreover, in the NGPSI approach, the aspiration levels were defined based on the meaning of each initial indicator. It would be interesting to fix those levels through the input of an expert panel, and then compare those results to information from the current assessment. Achieving this added tier of precision would require participation of stakeholders and decision makers in the process of evaluating sustainability Fig. 4.

5. Conclusions Performance indicators (PIs) are widely used by water regulators to establish benchmarks to control the quality of service and/or to set water tariffs. However, PIs do not provide a holistic assessment of the general performance of water companies since the PI system is comprised of a large number of indicators. To overcome these limitations, PIs should be quantitatively aggregated,

converting them into a synthetic indicator which provides a multidimensional assessment of water company performance. There is a growing interest in improving both the short- and long-term sustainability of water companies. However, most of the previous studies have focused only on one dimension of sustainability, i.e., they have evaluated the economic, social or environmental aspects individually, without integrating multiple parameters into a single, synthetic metric that can be applied to water company sustainability. This paper presents two methodological approaches to evaluate sustainability, embracing social, economic and environmental PIs as part of the process. The first approach is the distanceprincipal component which is a statistical MCDA based on PCA and the distance to a reference point. The second method presented is a non-statistical MCDA, such as global programming synthetic indicator. Our data show that both techniques are very useful in assessing the sustainability of water companies holistically since they integrate the set of initial PIs into a synthetic, composite indicator which incorporates key qualitative and quantitative metrics. An empirical application was developed to assess the sustainability of a sample of 154 Portuguese water companies. The initial set of PIs was composed by 14 indicators categorized as economic, social or environmental. Using the two methods, a synthetic indicator of sustainability was computed for each company. This approach allows for normalized comparison of company performance with regard to sustainability. Results illustrated that both methods yield similar water company rankings, confirming

586

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587

the utility of each and supporting the proposition that the synthetic indicator computed for each water company is reliable and, therefore, an appropriate tool for facilitating the decision-making process. From a managerial and policy perspective, the methods and results presented in this study are of key importance. There is ample evidence that a significant number of water companies have room to improve their short- and long-term sustainability, a task that can be supported and directed through use of the integrated indicator described here. In addition, stakeholders within the water industry can use our findings in this benchmarking exercise to raise social, environmental and economic awareness of the sustainability issues, and place appropriate pressure on water companies to work constructively toward sustainable goals. The development of a synthetic indicator provides a tool for ranking water companies according to their respective sustainability. This information is essential in regulated water industries to develop and implement incentives and regulatory measures designed to improve the sustainability of urban water services. Acknowledgement María Molinos-Senante wish to acknowledge the financial assistance received from the Conicyt through the program Fondecyt postdoctorado (3150268). Fátima Pérez, Trinidad Gómez and Rafael Caballero would also like to Thank The Junta de Andalucía for financial support (SEJ-417) Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ecolind.2015. 10.009. References Ahn, J., Kang, D., 2014. Optimal planning of water supply system for long-term sustainability. J. Hydro-Environ. Res. 8 (4), 410–420. Aydin, N.Y., Mays, L., Schmitt, T., 2014. Sustainability assessment of urban water distribution systems. Water Res. Manage. 28 (12), 4373–4384. Balkema, A.J., Preisig, H.A., Otterpohl, R., Lambert, F.J.D., 2002. Indicators for the sustainability assessment of wastewater treatment systems. Urban Water 4 (2), 153–161. Bana e Costa, C.A., Vansnick, J.-C., 1994. MACBETH – an interactive path towards the construction of cardinal value functions. Int. Trans Oper. Res. 1 (4), 489–500. Blanc, I., Friot, D., Margni, M., Jolliet, O., 2008. Towards a new index for environmental sustainability based on a DALY weighting approach. Sustain. Dev. 16 (4), 251–260. Blancas, F.J., González, M., Lozano-Oyola, M., Pérez, F., 2010a. The assessment of sustainable tourism: application to Spanish coastal destinations. Ecol. Indic. 10 (2), 484–492. Blancas, F.J., Caballero, R., González, M., Lozano-Oyola, M., Pérez, F., 2010b. Goal programming synthetic indicators: an application for sustainable tourism in Andalusian coastal countries. Ecol. Econ. 69 (11), 2158–2172. Blancas, F.J., Lozano-Oyola, M., González, M., Guerrero, F.M., Caballero, R., 2011. How to use sustainability indicators for tourism planning: the case of rural tourism in Andalusia (Spain). Sci. Total Environ. 412–413, 28–45. Boscheck, R., Clifton, J.C., Díaz-Fuentes, D., Oelmann, M., Czichy, C., Alessi, M., Treyer, S., Wright, J., Cave, M., 2013. The regulation of water services in the EU. Intereconomics 48 (3), 136–158. Caballero Fernandez, R., Cruz Morato, M.A., 2011. Análisis de la exclusión social en la Unión Europea basado en indicadores sintéticos. Recta 12 (11), 85–104. Carvalho, P., Marques, R., 2015. Estimating size and scope economies in the Portuguese water sector using the most appropriate functional form. Eng. Econ., forthcoming (DOI:10.1080/0013791X.2013.873507). Cashman, A., Lewis, L., 2007. Topping up or watering down? Sustainable development in the privatized UK water industry. Bus. Strateg. Environ. 16 (2), 93–105. Corder, G.W., Foreman, D.I., 2014. Nonparametric Statistics: A Step-by-Step Approach. Wiley, ISBN 978-1118840313. Da Cruz, N., Marques, R., 2012. Mixed companies and local governance: no man can serve two masters. Public Admin. 90 (3), 737–758. De Witte, K., Marques, R.C., 2012. Gaming in a benchmarking environment. A nonparametric analysis of benchmarking in the water sector. Water Policy 14 (1), 45–66.

Duarte, A.A.L.S., Rodrigues, G.M.C., Ramos, R.A.R., 2009. A global service quality index to evaluate the performance and sustainability in water supply utilities. WSEAS Trans. Environ. Dev. 5 (12), 759–769. Diaz-Balteiro, L., Romero, C., 2004. Sustainability of forest management plans: a discrete goal programming approach. J. Environ. Manage. 71 (4), 351–359. EPA, 2012. Planning for Sustainability. A Handbook for Water and Wastewater Utilities. EPA, Seattle. Foxon, T.J., McIlkenny, G., Gilmour, D., Oltean-Dumbrava, C., Souter, N., Ashley, R., Butler, D., Jowitt, P., Moir, J., 2002. Sustainability criteria for decision support in the UK water industry. J. Environ. Plan. Manage. 45 (2), 285–301. Giannetti, B.F., Bonilla, S.H., Silva, C.C., Almeida, C.M.V.B., 2009. The reliability of experts’ opinions in constructing a composite environmental index: the case of ESI 2005. J. Environ. Manage. 90 (8), 2448–2459. Hamouda, M.A., Nour El-Din, M.M., Moursy, F.I., 2009. Vulnerability assessment of water resources systems in the Eastern Nile Basin. Water Res. Manage. 23 (13), 2697–2725. Hernández-Sancho, F., Molinos-Senante, M., Sala-Garrido, R., Del Saz-Salazar, S., 2012. Tariffs and efficient performance by water suppliers: an empirical approach. Water Policy 14 (5), 854–864. Jiang, M.-X., Zhang, J.-X., Li, M., 2013. Robust online object tracking based on feature grouping and 2DPCA. Math. Probl. Eng., article ID: 352634. Klostermann, J.E.M., Cramer, J., 2007. Social construction of sustainability in water companies in the Dutch coastal zone. J. Cleaner Prod. 15 (16), 1573–1584. Lozano-Oyola, M., Blancas, F.J., González, M., Caballero, R., 2012. Sustainable tourism indicators as planning tools in cultural destinations. Ecol. Indic. 18, 659–675. Lundie, S., Peters, G., Ashbolt, N., Lai, E., Livingston, D., 2010. A sustainability framework for the Australian water industry. Water 33 (7), 83–88. Marques, R.C., 2006. A yardstick competition model for Portuguese water and sewerage services regulation. Util. Policy 14 (3), 175–184. Marques, R., 2008. Comparing private and public performance of Portuguese water services. Water Policy 10 (1), 25–42. Marques, R., Berg, S., 2011. Public–private partnership contracts: a tale of two cities with different contractual arrangements. Public Admin. 89 (4), 1585–1603. Marques, R., Simões, P., 2008. Does the sunshine regulatory approach work? Governance and regulation model of the urban waste services in Portugal. Res. Conserv. Recycl. 52 (8/9), 1040–1049. Marques, R.C., Simões, P., Pires, J.S., 2011. Performance benchmarking in utility regulation: the worldwide experience. Pol. J. Environ. Stud. 20 (1), 125–132. Marques, R.C., da Cruz, N.F., Pires, J., 2015. Measuring the sustainability of urban water services. Environmental Science and Policy 54, 142–151. Molinos-Senante, M., Gomez, T., Garrido-Baserba, M., Caballero, R., SalaGarrido, R., 2014. Assessing the sustainability of small wastewater treatment systems: a composite indicator approach. Sci. Total Environ. 497, 607–617. Molinos-Senante, M., Sala-Garrido, R., Lafuente, M., 2015. The role of environmental variables on the efficiency of water and sewerage companies: a case study of Chile. Environ. Sci. Pollut. Res. 22 (13), 10242–10253. Nardo, M., Siasana, M., Saltelli, A., Tarantola, S., 2005. Tools for Composite Indicators Building. Institute for the Protection and Security of the Citizen, European Commission. Palme, U., Tillman, A.-M., 2008. Sustainable development indicators: how are they used in Swedish water utilities? J. Cleaner Prod. 16, 1346–1357. Palme, U., Tillman, A.-M., 2009. Sustainable urban water systems in indicators: researchers’ recommendations versus practice in Swedish utilities. Water Policy 11 (2), 250–268. Pérez, V., Guerrero, F., González, M., Pérez, F., Caballero, R., 2013. Composite indicator for the assessment of sustainability: the case of Cuban nature-based tourism destinations. Ecol. Indic. 29, 316–324. Perez, M., Tujchneider, O., Paris, M., D’Elía, M., 2015. Sustainability indicators of groundwater resources in the central area of Santa Fe province, Argentina. Environ. Earth Sci. 73 (6), 2671–2682. Pinto, F., Da Cruz, N., Marques, R., 2015. Contracting water services with public and private partners: a case study approach. J. Water Supply: Res. Technol. – AQUA 64 (2), 194–210. Reisi, M., Aye, L., Rajabifard, A., Ngo, T., 2014. Transport sustainability index: Melbourne case study. Ecol. Indic. 43, 288–296. Rojas-Torres, M.G., Nápoles-Rivera, F., Ponce-Ortega, J.M., Serna-González, M., ElHalwagi, M.M., 2014. Optimal design of sustainable water systems for cities involving future projections. Comput. Chem. Eng. 69, 1–15. Saaty, T.L., 2008. Decision making with the analytic hierarchy process. Int. J. Serv. Sci. 1 (1), 83–98. Sahely, H.R., Kennedy, C.A., Adams, B.J., 2005. Developing sustainability criteria for urban infrastructure systems. Can. J. Civil Eng. 32 (1), 72–85. Salvati, L., Carlucci, M., 2015. Towards sustainability in agro-forest systems? Grazing intensity, soil degradation and the socioeconomic profile of rural communities in Italy. Ecol. Econ. 112, 1–13. Schulz, M., Short, M.D., Peters, G.M., 2012. A streamlined sustainability assessment tool for improved decision making in the urban water industry. Integr. Environ. Assess. Manage. 8 (1), 183–193. Simões, P., Marques, R., 2012. Influence of regulation on the productivity of waste utilities. What can we learn with the Portuguese experience? Waste Manage. 32 (6), 1266–1275.

M. Molinos-Senante et al. / Ecological Indicators 61 (2016) 577–587 Singh, R.K., Murty, H.R., Gupta, S.K., Dikshit, A.K., 2012. An overview of sustainability assessment methodologies. Ecol. Indic. 15 (1), 281–299. Singh, M.R., Upadhyay, V., Mittal, A.K., 2010. Addressing sustainability in benchmarking framework for Indian urban water utilities. J. Infrastruct. Syst. 16 (1), 001001QIS, 81–92. Somarriba, N., Pena, B., 2009. Synthetic indicators of quality of life in Europe. Soc. Indic. Res. 94 (1), 115–133. Tabesh, M., Saber, H., 2012. A prioritization model for rehabilitation of water distribution networks using GIS. Water Res. Manage. 26 (1), 225–241.

587

Voces, R., Diaz-Balteiro, L., Romero, C., 2012. Characterization and explanation of the sustainability of the European wood manufacturing industries: a quantitative approach. Expert Syst. Appl. 39 (7), 6618–6627. WCED (World Commission on Environment Development), 1987. Report of the World Commission on Environment and Development: Our Common Future. Wong, S., Pang, L., 2003. Motivators to creativity in the hotel industry – perspectives of managers and supervisors. Tourism Manage. 24, 551–559.