Social multicriteria evaluation of conflict over rural electrification and solar energy in Spain

Environment and Planning C: Government and Policy 2008, volume 26, pages 712 ^ 727

doi:10.1068/c06105s

Social multicriteria evaluation of conflict over rural electrification and solar energy in Spain Giuseppe Munda

European Commission Joint Research Centre, Institute for the Protection and Security of the Citizen (IPSC), TP 361, 21020 Ispra (VA), Italy and Department of Economics and Economic History and Institute for Environmental Sciences and Technologies, 08193 Bellaterra, Barcelona, Spain; e-mail: [email protected]

Daniela Russi }

Department of Economics and Economic History and Institute for Environmental Sciences and Technologies, 08193 Bellaterra, Barcelona, Spain; e-mail: [email protected] Received 8 March 2006; in revised form 26 March 2007; published online 9 May 2008

Abstract. A case study is presented with two objectives: (i) to give a clear and simple illustration of social multicriteria evaluation in the field of rural renewable-energy policy, and (ii) to help in understanding the extent to which, and the circumstances under which, solar energy is suitable for electrifying isolated farmhouses. Our aim is to provide public decision makers with insight into the conditions that favour the diffusion of renewable energy. This should facilitate the design of more effective energy policies for rural communities.

1 Introduction According to a World Bank study, of 3.3 billion people living in rural areas, only 1.5 billion have access to electricity (Cabraal et al, 1996). Lack of electrification in rural areas is an especially important issue in Southern countries (Chaurey et al, 2002), but it is also important in Europe (Vallve¨ and Serrasolses, 1997). For example, in Catalonia, 1063 households still lack electricity (ICAEN, 2002). The reason for this in isolated areas is the high cost of extending the grid (Gabler, 1998). Moreover, in forested areas, grid extension implies deforestation, risk of fire, possible damage to avifauna, and an impact on the landscape. In this context, solar energy may represent a viable alternative to traditional rural electrification. Niche markets, where renewable energy is already competitive or almost competitive with traditional energy, such as stand-alone photovoltaic (PV) applications (Hoffmann, 2005), may help to expand the solar-energy market by increasing, through scale economies and investments in research, its competitiveness (Masini and Frankl, 2002). However, the state is an indispensable actor for launching renewable energies. Solar energy needs to be subsidised while rural electrification requires engaging various private and public actors with different, and possibly conflicting, values, interests, and requirements. Social multicriteria evaluation (SMCE) can be a useful policy framework to guarantee that decisions on energy policies for SMCE rural areas are made as transparent as possible, and to guarantee that all the involved actors can participate (Munda, 2004). We present a case study to deal with many issues that are typical of the rural electrification problem. Tagamanent is a village located in Montseny Natural Park, near Barcelona, and is where a conflict arose in 1994 over how to provide some isolated farmhouses with electricity. In order to understand the reasons for the conflict we perform a retroactive SMCE, with two objectives: first to give a clear and simple illustrative example of the application of SMCE in the field of renewable-energy } Corresponding author.

Evaluation of conflict over rural electrification and solar energy in Spain

713

policies; and, second, to help to understand the extent to which and under which circumstances solar energy is suitable for electrifying isolated farmhouses. The next section introduces SMCE, outlining why it is a useful methodology to support public policies. We then show, step by step, how the SMCE methodology is adapted to the Tagamanent case study. Our conclusions reveal how SMCE can cut through the complexity of a social conflict and can reveal key issues underlying an ongoing problem. 2 SMCE for addressing complexity The real world is characterised by deep complexity. This obvious observation has important implications for the manner in which policy problems are represented and how decision making is framed. Any representation of a complex system reflects only a subset of its possible representations (Giampietro et al, 2006). A consequence of these deep subjectivities is that, in any normative exercise connected to a public-decision problem, one has to choose an operational definition of `value'. This is in spite of the fact that social actors with different interests, cultural identities, and goals may have different definitions of value. That is, to reach a ranking of policy options there is a prior need for deciding what is important for different social actors as well as what is relevant for the representation of the real-world entity described in the model (Funtowicz et al, 2002; Munda, 2004). In empirical evaluations of public policies and publicly provided goods, multicriteria-decision analysis seems to be an appropriate policy tool because it can take into account a wide range of assessment criteria (Munda, 2005). Also, the management of a policy process involves many layers and kinds of decisions, and requires the construction of a dialogue process among many social actors: individual and collective, formal and informal, local and nonlocal. An outcome of this discussion is that the political and social framework must find a place in evaluation exercises. This is the objective of SMCE (Munda, 2004), which can overcome the pitfalls of a technocratic approach by applying different methods of sociological research [eg institutional analysis, as in Corral Quintana (2000)]. By means of interviews and focus groups it is possible to ascertain people's desires and it is then possible to develop a set of policy options and evaluation criteria (De Marchi et al, 2000).(1) In operational terms, the application of an SMCE framework involves seven main steps (Munda, 2005): (i) identification of relevant social actors, by means of institutional analysis; (ii) definition of social actors' values, desires, and preferences, mainly through in-depth interviews and focus groups; (iii) generation of policy options and evaluation criteria;(2) (iv) construction of the multicriteria impact matrix, which synthesises the scores of all criteria for all alternatives (Janssen and Munda, 1999);(3) (v) construction of an equity impact matrix using NAIADE (Joint Research Centre, 1996; Munda, 1995), which allows representation of the distance between the positions of the social actorsöthat is, the degree of conflict and the possible coalitions amongst them; (vi) application of the multicriteria algorithm to obtain a final ranking of the available alternatives [numerous options are available, eg, Arrow and Raynaud (1986); (1) However, in SMCE, participation is used as an input for the analysis, but criteria and weights are not directly derived from participation such as in participatory multicriteria analysis (Stagl, 2006) or deliberative multicriteria analysis (Proctor and Drechsler, 2006). An interesting discussion of pros and cons of these different methods can be found in Kallis et al (2006). (2) Criteria are indicators that assess the extent to which the different social actors' objectives are achieved by each alternative. (3) In this case, the equity analysis was not considered necessary because of the small number of social actors.

714

G Munda, D Russi

Figueira et al (2005); Munda (1995); Roy (1996); for a review of advantages and disadvantages of the different methodologies, see Mys­ iak (2006)]; (vii) sensitivity and robustness analysis of the final ranking. This last step is very important due to the unavoidable degree of uncertainty that characterises most real-world decision-making processes (Saltelli et al, 2004). 3 SMCE in Tagamanent village, Montseny Natural Park Montseny Natural Park (which has an area of 301 km2 and a population of almost 1000), a UNESCO Biosphere Reserve since 1978, is situated in northern Catalonia, between Girona and Barcelona. It is only 40 km from Barcelona, and is a popular place for outdoor excursions (Boada and Junca©, 2002). Most farmhouses inside the park were built centuries ago, by carbon¬eros (charcoal makers), farmers, and stockbreeders, and constitute an important architectonic heritage. Farmhouse abandonment constitutes a particularly serious problem. However, in the last few years a moderate repopulation process has taken place, driven by people coming from the cities. There has also been an expansion of the tertiary sector (mainly restaurants and pensions) inside the park. There is a recognised lack of electricity in the area. In order to solve the electrification deficit, in 1994 the Park administration, Servei de Parcs Natural (SPN), launched an electrification plan by means of PV SPN systems. Between 1995 and 2000 this provided thirty-two isolated farmhouses with electricityö almost 30% of the permanently inhabited farmhouses (Argemi and Serrasolses, 2002). However, in Tagamanent the desire was for grid extension, rather than PV, to supply electricity. Most farmhouse and park inhabitants disagreed with the decision in favour of solar energy. SMCE was applied to the present case study with the aim of taking into account the inherent complexity of the issueöthat is, not only dealing with economic costs and technical performance but also employing other incommensurable criteria reflecting the objectives and the interests of the social actors. SMCE was used retrospectively, in 2003, after the park administration had decided in favour of solar energy. This ex post analysis, in explaining the positions and the choices of the social actors, as well as explaining the different impacts of the policy options, constitutes a useful basis for the ongoing public debate. Insight is also gained as to the factors that affirm the solarenergy choice. Retrospective studies are still quite unusual in public-policy analysis, but they might constitute a useful instrument to increase public accountability of the decision-making processes. 3.1 Institutional analysis

The first step of an SMCE is the institutional analysis, which gathers information on the institutional structure and on the social actors. Documentary evidence was obtained from the archives of the park and Tagamanent municipality. In addition, we conducted open, anonymous, in-depth interviews with various stakeholders.(4) The Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms (SEBA), a nonprofit-making association that organises the installation of off-grid PV systems in isolated households, installed the equipment and, in exchange for a monthly fee of about 20, committed itself to providing users with technical supervision, insurance, and free maintenance. (4) We carried out fifteen interviews (four telephonically and the rest face to face) between May 2003 and February 2004 with social actors: two with the Mayor, one with the SPN technician in charge of rural electrification, eight with farmhouse owners, two with farmhouse leaseholders, and two with people running a restaurant in a farmhouse belonging to the park administration. Also, two interviews with Mart|¨ Boada, a natural scientist who has been studying Montseny for decades, were essential for getting to know the history and the general characteristics of the massif as well as the strategy of the Park administration.

Evaluation of conflict over rural electrification and solar energy in Spain

715

SPN paid 45% of the expense, and SEBA financed 34% using subsidies from the Spanish and the Catalan governments and from the EU. However, in Tagamanent (with 235 inhabitants and an area of 43.48 km2 ) PV was never accepted and the programme reinitiated the debate over rural electrification, which had been an issue since 1993. In 1996 SPN rejected an electrification plan proposed by the Tagamanent municipality and prepared by Fecsa, the electricity company, because of its environmental impact. Instead, SPN commissioned SEBA to carry out a study to compare different options for rural electrification for Tagamanent (Trama Tecno Ambiental, 1998). In the following years PV systems were installed in seven out of the twenty-four scattered farmhouses.(5) However, the amount of installed PV power was lower (7.4 kWp for the seven households) than that necessary to supply enough electricity in the winter months. Even after the installation of the PV systems, the conflict between the Mayor, supported by most farmhouse owners (in favour of grid extension), and SPN (in favour of PV) remained unresolved. In various heated meetings the Mayor unsuccessfully tried to convince SPN to support electric-line extension instead of PV. There are four main social actors. First is the SPN, tasked with protection of the Montseny environment. SPN was interested in the electrification of the isolated farmhouses because that would promote the repopulation of the park, but only by means of PV systems. This position has been very firm during the last ten years. Second are the owners of the farmhouses who mostly use them as weekend residences or who rent them out. Some use the land as pasture for their cattle, which they entrust to local breeders. Owners prefer `traditional' electricity, because of long-term benefits such as the revaluation of the farmhouses and the possibility of running a restaurant or a rural pension. In contrast, PV systems have a limited lifetime, and must be periodically replaced. Third are inhabitants,(6) mostly from the cities, who are leasing farmhouses as residences.(7) These people have chosen to live in the park to be closer to nature, but do not live on primary activities. The inhabitants want reliable energy at a reasonable cost and in a sufficient quantity, especially if they are running an enterprise. Fourth is the Mayor of Tagamanent. He claims that grid extension would encourage the repopulation of the park, while PV is unable to supply enough energy to support economic activities. In other words, PV had a very high opportunity cost. In the remainder of this study this position is treated as being identical to that of farmhouse owners. In fact, the decision on rural electrification involved three stages. First, SPN had to decide whether to allow and partly finance PV and/or grid extension (it went for solar energy). Second, if SPN had allowed and made both traditional and solar energy affordable, the farmhouse owners would have had to decide between the two alternatives. Third, if the owners refused to pay the cost of electrification, the inhabitants could still decide to pay themselves for one of the two electrification modalities. 3.2 Policy options

We analysed three alternatives for the fourteen households to be electrified. (8) First was the initial Fecsa projectöelectric-grid extension in one single stretch. This includes 12.2 km of middle voltage (25 kV) and 3 km of low voltage (380 V) line. Second was the proposal by SEBA that the electric grid be extended by means of two stretches, and (5) Most

of the nonelectrified farmhouses were in ruins, or were abandoned. included in the inhabitant category the people running a restaurant (rented from the park administration) because their interests could be assimilated to those of the inhabitants. (7) Only two of the farmhouses were inhabited by the owners. (8) The farmhouses in a ruinous state, that were not planned to be rehabilitated, were not considered. (6) We

716

G Munda, D Russi

with some environmental measures (Trama Tecno Ambiental, 1998).(9) Third was the electrification by means of stand-alone PV systems, which involves the following assumptions: the energy requirement would be that of an average Spanish household in 1998,(10) in order to set a level playing field by assuming that the people in the park would require the same amount of electricity as people living in cities. In order to allow for the worst-case scenario (11) we calculated an installed power of around 2.9 kWp per household (40.6 kWp in total), even though, for budget reasons, only a severely limited number of modules were then really installed (7.4 kWp in total), leading to a much less satisfactory performance of the PV option than theoretically possible. Finally, we took a twenty-year time horizon, in order to allow for the replacement costs of PV components, assuming that the batteries would have to replaced after ten years, the solar modules and the structure would have to be replaced after twenty years, and the regulating machinery and inverters would have to be replaced after fifteen years. 3.3 Construction of multicriteria impact matrices

The third step of an SMCE is to select the criteria that will be used to evaluate the alternatives. Interviews with the social actors were used to establish alternatives and criteria. Then, we determined the criterion scores for each alternative using interviews with experts in different disciplines.(12) We built three multicriteria impact matrices, one for each group of social actors. Hence, instead of constructing a single impact matrix for the entire society and then constructing an equity matrix, as is normally done in SMCE (eg De Marchi et al, 2000, Gamboa and Munda, 2007), we dealt with distributional conflicts directly in the building of the impact matrices. The fact is that groups of social actors often have different points of view on a problem. The power structure in society determines the set of criteria (and therefore the final decision) that prevails. In the Tagamanent case, SPN was the most powerful social actor and was able to prevent one of the options öthe grid extension. In this sense, the retrospective SMCE increases the transparency and the public accountability of the decision-making process. In fact, citizens can go back to the criteria (and the objectives) that were considered important by the politicians who made the decision, and can debate them. The following three subsections describe how we estimated the impacts for each actor (SPN, the owners, and the inhabitants). The criteria are economic, environmental, and social. Environmental criteria could cover a range of issues (eg acidification, eutrophication, resource depletion) but were restricted to those of concern to the social actors with the exception of climate change, which is a major concern at all levels of the public administration.(13) The analysis is made more transparent by using qualitative impact criteria and, throughout, when doing so we assign one of three scores: high, low, or no impact. (9) SEBA suggested: (1) to bury some stretches of the line; (2) to extend the low-tension length and to reduce the medium-tension length; (3) not to use aerial transformers but to put them on the ground; and (4) to modify the itinerary in order not to affect the most environmentally sensitive areas. (10) 192 kWh/month per household ˆ 2 688 kWh/month in total (Ministerio de Industria y Energ|¨ a, 2000). (11) This is based on the electricity required during the winter season. We considered an average irradiance in winter to be equal to 1300 kWh/(m2yr) and a wary assumption of a 60% performance factor (accounting for power losses caused by atmospheric depositions, orientation, etc). (12) The experts interviewed were: a technician in charge of rural electrification for Fecsa; two PV installers of two different companies; an expert in forest fire prevention; an ornithologist who studies the damage to avifauna caused by electric lines; a specialist in life cycle assessment; and a biologist with experience in landscape protection. (13) Commitments made with the Kyoto Protocol require a combined effort of the local and national administrations.

Evaluation of conflict over rural electrification and solar energy in Spain

717

3.3.1 Servei de Parcs Naturals In order to evaluate the costs for society as a whole, we multiplied the required power (40.6 kWp) by the average cost of PV equipment, as indicated by SEBA (14) (around ( 18.3 per installed Wp (15) ), obtaining a total cost of around 743 000 for fourteen households. Also, we took into account the fee that SEBA asked for PV maintenance (around 240 per year) and we discounted the expenses that would have taken place in the future,(16) using equation (1):(17) X Cv Cb Cm Cs D ˆ Cb ‡Cs ‡Cm ‡ ‡ ‡ ‡ t ˆ 1, 2, ::: , 20, (1) T, 10 15 20 …1‡r† …1‡r† …1‡r† t …1‡r† where D is discounted cost, Cb , Cm , Cs represent the cost of batteries, the cost of regulating machinery and inverters, and the cost of solar modules, respectively; Cv is the annual SEBA fee; t is the period (twenty years); and r is the interest rate.(18) Using this calculation, the cost of PV panels for the fourteen households is 1193 000. Solar energy is more advantageous if consumption is low and the temporal horizon is short. For example, using the same assumptions and considering an electricity consumption of 84 kWh/month per household (as in the 1998 SEBA report) would have led to a total cost of 551 000, whereas, if the temporal horizon had been only ten years (ie without including the replacement cost), the total cost would have been 859 000. Combining the two latter assumptions, we would obtain a total cost of 355 000 instead of 1193 000. As regards the grid extension, we summed the cost of grid extension and the variable cost of electricity.(19) Calculated in this way, the total cost of Fecsa's original project would have been 731 000, whereas SEBA's proposal would have implied a cost of 796 000. Because SPN financed 45% of PV installation (but not the grid extension), the cost of SPN for rural electrification through PV would have been around 510 000, maintaining the assumptions explained above and including the replacement cost. As previously stated, SPN encouraged human presence in the park but hampered those activities that could potentially jeopardise the environment. Therefore, SPN (14) The price of PV equipment has been decreasing steadily since. Today, the average module retail price, which represents approximately 50ö60% of a complete PV system (excluding batteries), can be estimated at around 5 ^ 6/Wp (see http://www/solarbuzz.com/Moduleprices.htm). (15) All the costs indicated in this paper include value added tax (VAT). (16) We discounted the costs using SWAP rates, which are used by financial entities for discounting costs and benefits of private investments. They are formulated by combining the actual interest rate with forecasts of future economic performance. SWAP rates increase over time because they assume uncertainty increases with time. They derive their name from the practice of swapping between financial entities loans at fixed and floating rates. The result was different discount rates for the costs taking place in each year over twenty years, with the base year being 1998. The rate ranged between 3.1% in 1999 and 4.7% in 2017. (17) We made some strong assumptions: (1) the cost of PV components would not change in the following twenty years. This assumption certainly overestimated the price of replacement of PV components, because the price of PV elements is quickly decreasing; (2) SEBA's fee and the consumption of an average household would not change; (3) the solar systems' efficiency was held stable; and (4) we did not take into account either the expense for the electrogenerators, which were used in some farmhouses in order to complement solar energy, or the difference in the cost of energy-saving households appliances with respect to normal ones. (18) We estimated the costs of the PV components using information found in TFM (2002). (19) Obtained by multiplying the average consumption of Spanish households by the price of electricity in 1998 (Boletin Oficial de Espan¬a) N. 210, 27/12/1997, pages 8161 ^ 8168. The price under 15 kW (we assumed a contracted power of 4.4 kW) in 1998 was 257 pesetas/kW ( 1.54/kW) per month plus 14.61 pesetas/kWh ( 0.09/kWh). We added to these figures 16% VAT and some further taxes (around 5%), plus around 1 for the rental of the equipment.

718

G Munda, D Russi

chose the rural electrification modality that would have helped it to limit the potentially polluting activities inside the park. All interviewed social actors agreed on the fact that solar energy would have constituted a high limitation on enterprises. In contrast, traditional electricity would have provided a virtually infinite supply of energy, causing no limitation on economic activities. Environmental criteria are risk of fire, deforestation, risk for birds, and climate change. Risk of fire was the main reason for SPN opposing the electric grid, which could cause dangerous sparks.(20) According to the interviewed technician of the Catalan Forest Fire Prevention Service, the fire risk of the grid in Tagamanent would have been between `medium' and `high'. The two most frequent tree species in the park, oak and pine, are classified among the species that are very easily flammable the whole year round (Peix, 1999). According to the Catalan Department of Environmental Statistics, the area is one of high summer-fire risk. The criterion qualitative score that the Fecsa project would have entailed is high because the entire length of the electric line would have been aerial, whereas the SEBA proposal would have implied a low risk because part of the line would have been buried. PV panels have no forest-fire risk. Forest vegetation is an essential part of the park ecosystems, which, by statute, SPN had to protect. According to the Decree 268/1996,(21) corridors of 6 m and 2 m along low-tension and medium-tension lines, respectively, had to be free from vegetation. This meant that Fecsa's project would have implied a deforestation of 67 000 m2, whereas SEBA's proposal would have required deforesting only 57 000 m2 (actually, this was its main advantage). Stand-alone PV systems involve no deforestation. Electric lines can cause bird electrocution or collision of birds with the poles or the electric line. They are among the first causes of nonnatural death for many endangered species (Tinto¨ and Real, 2003). The ornithologist that we interviewed for this research affirmed that risk in Montseny forest would have been low because birds can use trees for perching, which reduces risk of electrocution. PV would have not affected avifauna. Using information on Spanish primary consumption and the equivalent carbon dioxide (CO2) emissions from each energy source (BUWAL, 1996; Houghton et al, 1994), we calculated that, for each kWh produced in Spain, approximately 0.5 kg of CO2 equivalent is emitted.(22) Multiplying this by the Spanish average domestic energy consumption gives 96 kg of CO2 equivalent emitted by traditional electricity. Alsema et al (1998) calculate that, for an off-grid PV system, the energy payback time(23) was at least seven years in 1998. Thus, assuming an overall lifetime of twenty years, the CO2 emissions caused by PV on a larger scale would have been 96  7/20 ˆ 34 kg of CO2 equivalent. Social criteria are education and landscape impacts.(24) One of the objectives of the Natural Park is to allow citizens to learn about ecosystems and traditional activities. PV panels are used in the park for educational purposes and visits to PV installations are often organised. PV panels have had a high educational effect via development of a consciousness among Montseny inhabitants and visitors about energy saving. (20) Between 1960 and 2002, 10% of all forest fires in Montseny Natural Park were caused by the grid (data from the Forest Fire Prevention Service, http://wwwmediambient.gencet.net/cat/el medi/ incendis). (21) Diari Oficial de la Generalitat de Catalunya N. 2236, 29 July 1996. (22) This figure is high because it refers to the entire life cycle, including extraction, transport, and transformation phases. (23) The period that it needs to supply the same amount of energy as that produced in terms of electricity. (24) Impact on landscape is a social criterion because it is a matter of social perception.

Evaluation of conflict over rural electrification and solar energy in Spain

719

The educational effect of traditional electricity is judged to be none. Electricity lines impact on the landscape, especially in mountainous and forest regions, because they remove trees to avoid fire risks and can be seen from a long distance as a yellow line zigzagging across the middle of the green forest. The poles have a visual impact, even though the rules inside the park would have required painting the poles green. Therefore, impact on landscape is rated high in the case of Fecsa's project and low in the case of SEBA's proposal, which reduces the grid visibility. The impact of PV on landscape is rated to be none. Table 1 indicates the impact matrix for SPN. Table 1. Impact matrix for Servei de Parcs Naturals (SPN). Dimension

Economic

Environmental

Social

Criteria

total cost cost for SPN risk of fire deforestation risk for birds emitted carbon dioxide (CO2) limitation to enterprises educational effect impact on landscape

Unit

1000 1000 qualitative 1000 m2 qualitative kg CO2 equivalent qualitative qualitative qualitative

Score Fecsa

SEBA

photovoltaic system

731 0 high 67 low 96

796 0 low 57 low 96

1328 570 none 0 none 34

none none high

none none low

high high none

Note: Fecsa is the electricity company and SEBA is the Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms.

3.3.2 Owners and inhabitants Economic criteria for owners concern the personal costs, the possibilities for enterprise development, and increased property values. Most users remarked very strongly during the interviews that cost is a crucial aspect for them. Costs are calculated taking into account that users need to cover 22% of the PV cost and 50% of the grid cost. The results for Fecsa's project, SEBA's proposal, and PV were 28 000, 31 000, and 21 000 per household, respectively. All social actors agree on the fact that one of the main reasons that could explain the hostility against PV was that it would have provided only a limited, and, to a certain extent, unpredictable amount of electricity, making the possibility of founding an enterprise based on solar energy `low'.(25) On the contrary, the possibility of founding an enterprise given by traditional electricity would have been high because it would have guaranteed a virtually infinite supply of energy. For the owners, the possibility of a revaluation of their properties was a good reason for rural electrification. Because the increase of the farmhouse value in monetary terms depends on many uncertain factors we preferred a qualitative evaluation. Therefore, the revaluation of the farmhouses would have been low for PV because their lifetime would have been limited and would have been high for traditional electrification, a long-lasting investment. Obviously, this criterion was important only for owners and not for leaseholders. Forest fire was the most important environmental risk for owners and inhabitants because it directly affected their property. Forest fire could cause economic damage (25) The restaurant and a small rural pension of Tagamanent functioned, in part, with solar energy but complemented it with electricity generators (which were very noisy and expensive).

720

G Munda, D Russi

and could jeopardise the farmhouses, the surrounding area, and even the safety itself of owners and inhabitants. The criterion scores were illustrated above. Social criteria are comfort and reliability. In rural areas, the feeling of isolation with respect to city dwellers is often strong, and this was mentioned in many interviews. For example, public transport was not provided inside the park, the paths to the farmhouses were often in bad condition, and there was no telephone grid. The park population sometimes felt that the Public Administration did not reward enough the fact that they contributed to maintaining the architectonic patrimony of the park. Owners and inhabitants of scattered farmhouses inside the park felt discriminated against in their energy use because of the higher price and the lower degree of comfort. As already mentioned, however, this latter issue was principally caused by the fact that only about 18% of the PV capacity theoretically necessary to guarantee the same amount of electricity required by typical Spanish households was installed. Had a full 40.6 kWp of PV been installed, the level of comfort that the owners and inhabitants would have enjoyed would have been comparable to the one possible by means of grid electricity. Since it was decided to carry out, here, the comparison between the two scenarios assuming the same level of availability of electricity, the same score was attributed to all alternatives under this criterion (ie high). Concern regarding blackouts is one of the factors that contributed to the hostility against PV systems. For those who run a restaurant or a pension this is a major problem because, when PV systems break down, clients get bad service. This criterion Table 2. Impact matrix for owners. Dimension

Economic

Environmental

Criterion

cost per household potential for founding a business revaluation of the farmhouses risk of fire

Unit

Score Fecsa

SEBA

photovoltaic system

1000 qualitative

28 high

31 high

23 low

qualitative

high

high

low

qualitative

high

low

none

Note: Fecsa is the electricity company and SEBA is the Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms. Table 3. Impact matrix for inhabitants. Dimension

Economic Environmental Social

Criterion

cost per household potential for founding a business risk of fire comfort reliability

Unit

Score Fecsa

SEBA

photovoltaic system

1000 qualitative

28 high

31 high

23 low

qualitative qualitative qualitative

high high high

low high high

none high low

Note: Fecsa is the electricity company and SEBA is the Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms.

Evaluation of conflict over rural electrification and solar energy in Spain

721

was considered only for the inhabitants, because the owners do not directly experience possible breakdowns. Solar-energy users affirmed that sometimes their PV equipment suffered from breakdowns. The solar-energy-system technicians confirmed that, from time to time, some of the PV components broke. PV reliability can be therefore considered to be low. According to two persons who have extended the grid to their farmhouses in Tagamanent, the grid breaks very seldom and its reliability is high. Tables 2 and 3 summarise the criteria scores for owners and inhabitants. 3.4 Aggregation

Desirable properties for a multicriteria ranking procedure in the framework of public policy and sustainability issues are discussed in Janssen and Munda (1999) and Munda (2005). In short, it is very important that such ranking methods are simple to guarantee consistency and transparency, and noncompensatory to avoid bad environmental or social consequences being systematically outperformed by good economic consequences, or vice versa; intensity of preference is not taken into account, thus avoiding compensability and allowing for weights to be importance coefficients and not to be trade-offs.(26) A simple ranking algorithm, taking these properties into account, is the Condorcet consistent rule [see Young and Levenglick (1978) for its social choice characterisation, and see Munda (2005) and Munda and Nardo (2007) for its implementation in a multicriterion framework]. Details of the procedure are given in the appendix. By applying this ranking algorithm, the outranking matrix described in table 4 is obtained. The maximum likelihood ranking of alternatives deriving from this outranking matrix is the one indicated in table 5. PV is the most preferred option for SPN; the Fecsa solution is the worst option. In fact, most of the criteria that were considered important for SPN belonged to the environmental dimension and were favourable to PV. By applying the same ranking algorithm to tables 2 and 3, the results reported in tables 6 and 7 are obtained. Table 4. Outranking matrix deriving from table 1 (Servei de Parcs Naturals).

Fecsa SEBA Photovoltaic

Fecsa

SEBA

Photovoltaic system

0 0.6 0.8

0.4 0 0.8

0.2 0.2 0

Note: Fecsa is the electricity company and SEBA is the Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms. Table 5. Maximum likelihood ranking of alternatives deriving from table 1 (Servei de Parcs Naturals). Photovoltaic Photovoltaic SEBA Fecsa SEBA Fecsa

SEBA Fecsa photovoltaic photovoltaic Fecsa SEBA

Fecsa SEBA Fecsa SEBA photovoltaic photovoltaic

2.2 1.9 1.6 1.4 1.1 0.8

Note: Fecsa is the electricity company and SEBA is the Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms. (26) Trade-off

weights show the intensity of preference and indicate how much of an advantage in a criterion is sufficient to compensate for a disadvantage in another criterion (for example, one might be willing to accept some environmental impact if it is compensated for by a sufficient economic income). Importance weights only indicate how important a criterion is, without referring to its criterion score.

722

G Munda, D Russi

Table 6. Maximum likelihood ranking of alternatives deriving from table 2 (owners). Fecsa SEBA photovoltaic SEBA Fecsa photovoltaic Fecsa photovoltaic SEBA SEBA photovoltaic Fecsa Photovoltaic Fecsa SEBA Photovoltaic SEBA Fecsa Note: Fecsa is the electricity company and SEBA is the Associacio¨ Auto©noms.

1.5 1.5 1.5 1.5 1.5 1.5 de Serveis Energe©tics Ba©sics

Table 7. Maximum likelihood ranking of alternatives deriving from table 3 (inhabitants). Fecsa SEBA photovoltaic SEBA Fecsa photovoltaic Fecsa photovoltaic SEBA SEBA photovoltaic Fecsa Photovoltaic Fecsa SEBA Photovoltaic SEBA Fecsa Note: Fecsa is the electricity company and SEBA is the Associacio¨ Auto©noms.

1.5 1.5 1.5 1.5 1.5 1.5 de Serveis Energe©tics Ba©sics

For the owners and inhabitants, the ranking of policy is the same öall three alternatives proved to be equally preferable. In some ways the advantages of PV (lower cost and lower environmental risk) are compensated for by the disadvantages (a smaller possibility of founding an enterprise, reduced revaluation for owners and reduced reliability for inhabitants). As regards the comparison between the Fecsa project and the modifications proposed by Seba (characterised by a higher cost and a lower environmental impact), the improvement in environmental protection assured by the latter were offset by an increase in costs. However, during the interviews, owners and inhabitants were not neutral between electric grid and PV. The reason is that, due to budget constraints, the amount of installed power was much lower than the one calculated here as being theoretically necessary.(27) Therefore, since the electricity provided by their PV systems was scarce they would have preferred the electric grid, which would not have imposed limitations on the electricity consumption, but only with the condition that part of the significantly higher cost would be subsidised. In this ranking exercise all criteria receive the same importance because no weight coefficient is used. This implies that the weight of each dimension depends on the number of criteria that belong to it. As a result, for SPN, more importance is given to the environmental dimension (five criteria out of nine) and, for owners, to the economic dimension (three criteria out of four). Let us, then, perform a sensitivity analysis by giving the same weight to all the dimensions used. The results are reported in tables 8, 9, and 10. The previous results are reinforced for SPN, for whom PV is again the best option, because both environmental and social criteria are in favour of solar energy. However, the ranking changes for owners and inhabitants because, (27) If the analysis had been performed while changing the assumptions for the PV scenario so as to reflect the actual installation (ie 7.4 kWp instead of 40.6 kWp), comfort would have been lower for PV than for the traditional electrification. All the other criteria would have remained unchanged. Even though the cost of PV would have been lower, since the method used here does not take into account the intensity of preference, the comparison among the alternatives according to the criterion `cost' would not have changed. In contrast, the criterion `comfort' would have assigned a better performance to the two grid options than to PV, leading to a possibly different ranking for the inhabitants than the one shown in table 7. Instead, the ranking for the owners would not have changed, because the criterion `comfort' was not considered to be important for them.

Evaluation of conflict over rural electrification and solar energy in Spain

723

Table 8. Sensitivity analysis for Servei de Parcs Natural's rankings. Photovoltaic Photovoltaic SEBA Fecsa SEBA Fecsa

SEBA Fecsa photovoltaic photovoltaic Fecsa SEBA

Fecsa SEBA Fecsa SEBA photovoltaic photovoltaic

2.2 1.9 1.6 1.4 1.1 0.8

Note: Fecsa is the electricity company and SEBA is the Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms. Table 9. Sensitivity analysis for owners' rankings. Photovoltaic SEBA Photovoltaic Fecsa SEBA Fecsa

SEBA photovoltaic Fecsa photovoltaic Fecsa SEBA

Fecsa Fecsa SEBA SEBA photovoltaic photovoltaic

2.0 1.7 1.7 1.3 1.3 1.0

Note: Fecsa is the electricity company and SEBA is the Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms. Table 10. Sensitivity analysis for inhabitants' rankings. Photovoltaic Potovoltaic SEBA Fecsa SEBA Fecsa

SEBA Fecsa photovoltaic photovoltaic Fecsa SEBA

Fecsa SEBA Fecsa SEBA photovoltaic photovoltaic

1.8 1.6 1.6 1.4 1.4 1.3

Note: Fecsa is the electricity company and SEBA is the Associacio¨ de Serveis Energe©tics Ba©sics Auto©noms.

by giving the same weight to each dimension rather than to each criterion, the overall relative weight of the environmental dimension (which is comprised here of the smallest number of criteria) increases and therefore PV becomes more preferable and Fecsa becomes the least preferable option. 4 Discussion and conclusions This study is an example of how an SMCE can be a means of addressing social conflict. The analysis began with a focus only on the energy issue. The problem that it intended to solve was the best way to electrify some isolated farmhouses in a natural park, taking environmental, social, and economic criteria into account. However, as the research continued, the interviews allowed understanding of various aspects of the problem that were not evident at first sight. Indeed, the rural electrification issue is not only a technical problem, but is also part of a larger political issue: the long-term political strategy for Montseny Natural Park. In order to explain this point, a distinction between technical and practical problems is helpful (Ravetz, 1971; Strand, 2002). The former can be solved using specialised technical knowledge, whereas the latter have to do with the objectives and the values of part of society, such as the wish for a clean environment, the wish for economic growth, or the wish for a fairer distribution of wealth. Thus, one can say that the solution of the technical problem (how to electrify the isolated farmhouses in Tagamanent) depended on how to give an answer to the practical

724

G Munda, D Russi

problem that is at the root of it: the conflict between different views on the development of the park. In fact, on one hand, SPN tends to adopt a conservative viewöto limit human interference with the natural ecosystems. Extending the grid to the isolated farmhouses would have had an impact on the environment and the landscape, and it might also have stimulated the foundation of enterprises inside the park. On the other hand, people living in or owning a farmhouse in Montseny thought that the policies encouraging the economic activities should be privileged and therefore would have preferred traditional electrification. This study shows that the relevant criteria, and hence the preferred alternative, are different for each group of social actors, and three different impact matrices are required. It might be argued that a debate should begin with who decides on park management, or, in other words, who can impose his or her set of evaluation criteria. Finally, some observations on the suitability of solar energy for rural electrification can be made. First, public incentives can greatly modify the preferences of the users. In the Tagamanent case, even though the PV systems were more expensive (the electricity consumption being equal), they were cheaper for the final users because of the subsidies. A lesson that can be learned is that the role of the public administration is crucial in promoting renewable energy. Second, the choice between solar and traditional energy depends on the temporal horizon. In fact, whereas, once installed, the electric grid provides electricity for a virtually infinite period of time, the PV components must be replaced from time to time. Therefore, solar energy appears to be preferable if the users are planning to stay in the farmhouse for a limited period of time (eg if they rent the house). Third, the activity carried out in the farmhouses is crucial. In fact, solar energy seems to be more appropriate for private households, whose energy consumption is low. On the contrary, people running an enterprise complain about low reliability and limited supply of solar energy. Finally, solar energy is diffuse and discontinuous in nature, and, for this reason, installation prices are high. If one wants to assure a certain level of service by means of off-grid PV, the dimension of the installation should be calculated under worst-case assumptions (winter months). Inevitably, in this way, stand-alone PV is a comparatively wasteful energy solution,(28) because it produces a surplus of electricity in summer, which cannot be stored in batteries. The fact that, in the Tagamanent case, the decision was made to install a limited amount of PV power clearly shows that the cost for PV was high and that the public administration could not afford to subsidise such a full-blown system. However, the price of PV components are rapidly decreasing, and a similar analysis could probably produce different results if performed in the near future. As a conclusion, in Tagamanent PV was the best option because, firstly, it was preferred by SPN due to environmental and social criteria and because, secondly, owners and inhabitants would have been neutral between PV and the two grid-extension options, if they had been provided with a sufficient amount of electricity. The conflict in Tagamanent may then be attributed to the fact that, due to budget constraints, not enough solar power was installed. Acknowledgements. Comments by Marti Boada, Joan Martinez-Alier, Marco Raugei, Clive L Spash, and two anonymous reviewers are gratefully acknowledged. Thanks are due to all persons representing different social actors, for their kind collaboration in the field work. This research was financially supported by the Catalan Agency for Management of University and Research Grants.

(28) On

the contrary, on-grid PV systems avoid energy wastes because they can be dimensioned according to the yearly average irradiation rate (instead of the lower winter rate), allowing users to buy electricity from the grid in winter and selling it back in the summer.

Evaluation of conflict over rural electrification and solar energy in Spain

725

It was also partly financed by the European Commission, research project: Development and Application of a Multi-Criteria Decision Analysis Software Tool for Renewable Energy Sources (MCDA RES), Contract NNE5-2001-273. References Alsema E A, Frankl P, Kato K, 1998,``Energy pay-back time of photovoltaic energy systems: present status and prospects'', paper presented at the Second World Conference on Photovoltaic Solar Energy Conversion, Vienna, http://www.chem.uu.nl/nws/www/publica/98053.pdf Argemi Relat J, Serrasolses J, 2002, ``Programa de electrificacio¨n rural en el Parque Natural del Montseny'', in Energ|¨ a, Sociedad y Medio Ambiente, Ed. Gobierno Vasco (Berekintza, S.L., Bilbao) pp 147 ^ 151 Arrow K J, Raynaud H, 1986 Social Choice and Multicriterion Decision Making (MIT Press, Cambridge, MA) Boada, Junca© M, 2002 El Montseny. Cinquanta Anys d'Evolucio¨ dels Paisatges (Publicacions de l'Abadia de Montserrat, Barcelona) Cabraal A, Cosgrove-Davies M, Schaeffer L, 1996, ``Best practices for photovoltaic household electrification programs: lessons from experiences in selected countries'', TP 324, World Bank, http://www.worldbank.org Chaurey A, Ranganathan M, Mohanty P, 2002,``Electricity access for geographically disadvantaged rural communitiesötechnology and policy insights'' Energy Policy 6 23 Corral Quintana S A, 2000 Una Metodolog|¨ a Integrada de Exploracio¨n y Comprensio¨n de los Procesos de Elaboracio¨n de Pol|¨ ticas Pu¨blicas PhD dissertation, Department of Economics, University of La Laguna, Tenerife De Marchi B, Funtowicz S O, Lo Cascio S, Munda G, 2000, ``Combining participative and institutional approaches with multi-criteria evaluation: an empirical study for water issue in Troina, Sicily'' Ecological Economics 34 267 ^ 282 Figueira J, Greco S, Ehrgott M (Eds), 2005 Multiple-criteria Decision Analysis: State of the Art Surveys (Springer, New York) Funtowicz S O, Martinez-Alier J, Munda G, Ravetz J, 2002, ``Multicriteria-based environmental policy'', in Implementing Sustainable Development Eds H Abaza, A Baranzini (UNEP/Edward Elgar, Cheltenham, Glos) pp 53 ^ 77 Gabler H, 1998, ``Autonomous power supply with photovoltaics: photovoltaics for rural electrificationöreality and vision'' Renewable Energy 5 512 ^ 518 Gamboa G, Munda G, 2007, ``The problem of wind-park location: a social multi-criteria evaluation framework'' Energy Policy 35 1564 ^ 1583 Giampietro M, Mayumi K, Munda G, 2006, ``Integrated assessment and energy analysis: quality assurance in multi-criteria analyses of sustainability'' Energy 31 59 ^ 86 Hoffman W, 2005, ``PV solar electricity in Europeöcompeting with Japan, USA and SEA'', http://www.epia.org/03DataFigures/Presentations/DG Research 200511211.ppt Houghton J T, Meira Filho L G, Bruce J, Lee H, Callander B A, Haites E, Harris N, Maskell K (Eds), 1994 Climate Change 1994: Radiative Forcing of Climate Change and an Evaluation of the IPCC IS92 Emission Scenarios (Cambridge University Press, Cambridge) ICAEN, 2002, ``Plan de l'energia a Catalunya en l'horitzo¨ de l'any 2010'', Institut Catala d'Energia, Generalitat de Catalunya, Departament d'Indu¨stria, Comerc° i Turisme Janssen R, Munda G, 1999, ``Multi-criteria methods for quantitative, qualitative and fuzzy evaluation problems'', in Handbook of Environmental and Resource Economics Ed. J van den Bergh (Edward Elgar, Cheltenham, Glos) pp 837 ^ 852 Joint Research Centre, 1996, ``NAIADE manual översion 1.0.ENG'', Joint Research Centre of the European Commission, Ispra, Italy, http://www.aiaccproject.org/meetings/Trieste 02/ trieste cd/software/NAIADE/naiade.PDF Kallis G,Videira N, Antunes P, Guimara¬es Pereira A, Spash C L, Coccossis H, Corral Quintana S, del Moral L, Hatzilacou D, Lobo G, Mexa A, Paneque P, Pedregal Mateos B, Santos R, 2006, ``Participatory methods for water resources planning'' Environment and Planning C: Government and Policy 24 215 ^ 234 Masini A, Frankl P, 2002, ``Forecasting the diffusion of photovoltaic systems in southern Europe'' Technological Forecasting and Social Change 70 39 ^ 65 Ministero de Industria Energ|¨ a, 2000, ``Estad|¨ stica de la industria de energ|¨ a ele¨ctrica 1998'', Ministerio de Industria, Turismo y Comercio, Madrid, http://www.mityc.es/Balances/Seccion/ Publicaciones/Estad|¨ sticas/Electricas/ElectricasAnuales/ Munda G, 1995 Multicriteria Evaluation in a Fuzzy Environment (Physica, Heidelberg)

726

G Munda, D Russi

Munda G, 2004, ``Social multi-criteria evaluation: methodological foundations and operational consequences European Journal of Operational Research 158 662 ^ 677 Munda G, 2005,``Multi-criteria decision analysis and sustainable development'', in Multiple-criteria Decision Analysis. State of the Art Surveys Eds J Figueira, S Greco, M Ehrgott (Springer, New York) pp 953 ^ 986 Munda G, Nardo M, 2007, ``Non-compensatory/non-linear composite indicators for ranking countries: a defensible setting'' Applied Economics (forthcoming) Mys­ iak J, 2006,``Consistency of the results of different MCA methods: a critical review Environment and Planning C: Government and Policy 24 257 ^ 277 Peix J (Ed.), 1999 Foc Verd II: Programa de Gestio© de Risc d'Incendi Forestal Departament d'Agricultura, Ramaderia i Pesca, Generalitat de Catalunya, Barcelona Proctor W, Drechsler M, 2006,``Deliberative multicriteria evaluation'' Environment and Planning C: Government and Policy 24 169 ^ 190 Ravetz J R, 1971 Scientific Knowledge and its Social Problems (Clarendon Press, Oxford) Roy B, 1996 Multicriteria Methodology for Decision Analysis (Kluwer, Dordrecht) Saltelli A, Tarantola S, Campolongo F, Ratto M, 2004 Sensitivity Analysis in Practice: A Guide to Assessing Scientific Models (John Wiley, New York) Stagl S, 2006, ``Multicriteria evaluation and public participation: the case of UK energy policy'' Land Use Policy 23 53 ^ 62 Strand R, 2002, ``Complexity, ideology and governance'' Emergence 4 164 ^ 183 TFM, 2002, ``Curs pra©ctic d'energ|¨ a solar fotovoltaica .3 Estudi econo©mic: ana©lisis comparatiu dels sistemes d'eletrificec|¨ o'', TFM Energ|¨ a Solar Fotovoltaica, Carrer Gaia©, nav 508110 Montcada i Reixac, Barcelona Tinto¨ A, Real J, 2003, ``Avaluacio¨ del risc d'electrocucio¨ d'aus en l|¨ nies ele©ctriques situades a la serralada pre-litoral de Barcelona'', Parc Natural de Sant Llorenc° del Munti i l'Obac i Aérees de Connexio¨ amb el Parc Natural del Montseny, report for Diputacio© de Barcelona and the Bosch and Gimpera Foundation, http://www.diba.es/porcsn/equipaments.asp?parc=4&tipus =7&m=42&s=259 Trama Tecno Ambiental, 1998, ``Estudi de necessitats energe©tiques al disseminat del municipi de tagamanent inclo©s dins del Parc Natural del Montseny'', Servei de Parcs Naturals, Diputacio¨ de Barcelona Vallve¨ X, Serrasolses J, 1997, ``Design and operation of a 50 kWp PV rural electrification project for remote sites in Spain'' Solar Energy 59 111 ^ 119 Young H P, Levenglick A, 1978, ``A consistent extension of Condorcet's election principle'' SIAM Journal on Applied Mathematics 35 285 ^ 300

Appendix Mathematical procedure for criterion aggregation

Given a finite set of criteria G ˆ fgm g, m ˆ 1, 2, ::: , M, and a finite set of alternatives A ˆ fan g, n ˆ 1, 2, ::: , N, let us assume that the evaluation of each alternative a n with respect to an evaluation criterion gm is based on an ordinal, or an interval, or a ratio scale of measurement. For simplicity of exposition, let us assume that a higher value of a criterion score is preferred over a lower one (the higher, the better). That is: aj Pak , gm …aj † > gm …ak † , aj Iak , gm …aj † ˆ gm …ak † ,

(A1)

where, P and I indicate a preference and an indifference relation, respectively, both fulfilling the transitive property (if ai Pak and ak Paj , then ai Paj ). Let us also assume the existence of a set of criterion weights W ˆ fwm g, m ˆ 1, 2, ::: , M, with M X wm ˆ 1 (A2) m ˆ1

Evaluation of conflict over rural electrification and solar energy in Spain

727

derived as importance coefficients. The mathematical problem to be dealt with then is how to use this available information to rank, in a complete preorder (ie without any incomparability relation(29)), all the alternatives from the best to the worst one. The mathematical aggregation convention can be divided into two main steps: 1. a pair-wise comparison of alternatives according to the whole set of criteria used; 2. ranking of alternatives in a complete preorder. An N  N matrix, E, called an outranking matrix (Arrow and Raynaud, 1986; Roy, 1996) can be built. Any generic element of E öej k , j 6ˆ köis the result of the pair-wise comparison, according to all the M criteria, between alternatives j and k. Such a global pair-wise comparison is obtained by means of equation (3): M   X 1 wm Pj k ‡ wm Ij k , (A3) ejk ˆ 2 mˆ1 where wm Pj k and wm Ij k are the weights of criteria presenting a preference and an indifference relation, respectively. It clearly holds that ej k ‡ ek j ˆ 1.

(A4)

The maximum likelihood ranking of alternatives is the ranking supported by the maximum number of criteria for each pair-wise comparison, summed over all pairs of alternatives considered. More formally, all the N(N ÿ 1) pair-wise comparisons compose the outranking matrix E. Call R the set of all N ! possible complete rankings of alternatives, R ˆ frs g, s ˆ 1, 2, .:: , N !. For each rs , compute the corresponding score js as  the summation of ej k over all the N 2 pairs j, k of alternatives. That is: P js ˆ ej k , (A5) where j 6ˆ k, s ˆ 1, 2, ::: , N ! and ej k 2 Rs. The final ranking (r ) is the one that maximises equation (5): P r , j ˆ max ej k , where ej k 2 R.

(29) The

relation between each pair of alternatives must be either of preference or indifference.

ß 2008 Pion Ltd and its Licensors

Conditions of use. This article may be downloaded from the E&P website for personal research by members of subscribing organisations. This PDF may not be placed on any website (or other online distribution system) without permission of the publisher.