دامنه انتخاب منبع و WTP برای برق در بازار نوظهور: نقش تجربه گذشته ، نگرانی های زیست محیطی و رفتار صرفه جویی در انرژی

This study analyzes customers' preferences and their willingness to pay (WTP) for certain service attributes in an electricity supplier choice context. Specifically, a stated preference choice experiment is conducted in the Canary Islands' residential market where limited competition exists. Preferences for different electricity suppliers and three level-of-service attributes are investigated, namely, supply reliability, share of renewable energies and availability of a complementary energy audit service. The results might be interpreted as an indication of different aspects new firms need to consider if they plan to enter in the market. There is an opportunity for new companies to establish in the market, though evidence of brand loyalty to the current company and/or significant switching costs are also found, especially in the case of older people. Regarding the estimated WTP, several results should be highlighted. First, customers who have experienced more serious outages in the past tend to show a higher WTP to reduce the outage frequency. Second, highly-educated respondents, those who state a great concern for the greenhouse gases (GHG) emissions and those who carry out energy saving actions in their homes exhibit a larger WTP for renewable energies. This empirical evidence provides useful information for authorities responsible for energy policy design.

In the last three decades, profound reforms have taken place in the electricity industry in the world. One of the most important features of these reforms has been the unbundling of different industry segments. The main purpose is to introduce competition into electricity generation and retail markets, and improve the efficiency of natural monopoly transmission and distribution stages. Whenever possible, the reforms also broke up the national distribution companies horizontally into several regional monopolies (see Green et al., 2006 and Hunt, 2002 among others). At the same time, several governments have started to include renewable energy on their public policy agenda. Recent initiatives, like the EU's 20-20-20 plan, have led many countries, including Spain, to establish as one of their main goals, an increase in the share of renewable energy in electricity generation to reduce greenhouse gases (GHG) emissions. In Spain, the enactment of Law 54/1997 on the Electricity Sector (LES) initiated a transition from a traditional regulatory regime to one that introduced competition in both generation and supply activities in the retail market. As happened in other countries, however, there has been no significant development of the retail market for residential customers (see Defeuilley, 2009) In December 2010, only 18.1% of domestic supply outlets, representing 27.5% of electricity consumption, were supplied through the free market. A large majority of users, thus, remain in the regulated market.1 Since 2010 electricity suppliers have begun to offer a series of complementary services or attributes in the supply of electricity, such as the joint sale of gas, various insurance offers, energy consulting, etc., in an effort to attract new clients or keep existing ones. The Canary Islands is a Spanish region, where the above mentioned trends have been manifested with certain specific features. On the one hand, the Canary Island electricity market is influenced by its status as a small and isolated electricity system and its emerging market condition, i.e. a market in transition from a traditional integrated and monopolistic one to a competitive deregulated one. There is almost no competition in the wholesale and retail market due to the system size and the vertical integration of the industry.2 On the other hand, the environment plays a key role in the region's economic activity, with tourism being the main economic industry (35% of GDP) and over 50% of its territory protected natural spaces. In fact, in 2006, the regional government approved an energy regulation document named Plan Energético de Canarias 2006 ( PECAN, 2006) with the aim of reducing the archipelago's energy sector's emission levels, and its dependence on oil with the introduction of natural gas and an increased reliance on renewable energy, especially solar and wind power. In this context, the knowledge of customers' preferences and of their willingness to pay (WTP) for the attributes that characterize the electricity supply can provide information about possible directions the current companies can take or what aspects new companies should consider to enter this emerging market. Likewise, eliciting WTP allows the evaluation of the market sustainability of government's goals in renewable energy generation or energy efficiency improvements, such as the ones contemplated in PECAN-2006. Residential customers' preferences for electricity suppliers and for several of their level-of-service attributes have already been analyzed in various studies. Stated choice experiments (Louviere et al., 2000) have been a widely used method for this purpose since the contributions by Cai et al. (1998), Goett (1998) and Revelt and Train (1998). Attributes like type of pricing, type of supplier, reliability, amount of renewables, and length of contract, among others have been found to be significant in explaining supplier choice. Nonetheless, there are reasons to suggest that supplier choice, and therefore the implicit WTP for different attributes, can be also influenced by additional factors. Cameron and Englin (1997) found evidence that WTP for environmental resources is systematically related to respondents' own experience with the good in question, where experience is interpreted as the number of years in which an individual has been a user of that resource. When choosing an electricity supplier, the experience as client of a company and the level of service experienced in the past (e.g. the importance of power outages suffered at home) could be aspects that need to be considered when investigating preferences. Other factors, such as lifestyle and awareness of environmental problems, have also shown to influence the criteria for energy consumption and result in greater WTP (see, for example, Brandon and Lewis, 1999). In particular, Seligman et al. (1979) and Uusitalo (1989) have argued that people's perceptions of their own contribution to energy problems is predictive of household energy conservation and that information designed to promote energy consciousness should emphasize such values. Analyzing this line of argument in greater depth raises the question of whether the value given to renewable energies is also related to household characteristics that are easily observable, such as energy-saving habits (e.g. the use of low energy consumption bulbs and domestic appliances, etc.), or to the level of concern about the environment. Knowing the influence of these factors on WTP will enable the assessment of the potential impact that fine-tuning policies related to the use of renewable energies, such as environmental awareness campaigns, might have. This study analyzes residential customers' preferences and their WTP for certain service attributes in an electricity supplier choice context in the retail market of the Canary Islands. In particular, preferences for the type of supplier and for three level-of-service attributes are investigated, namely, supply reliability, share of renewable energies and availability of a complementary energy audit service. Unlike previous studies on this subject, preferences for the electricity supplier are examined considering, on the one hand, how customers rate their current company to continue providing their electricity supply and, on the other hand, how customers rate the service provided by new companies. Furthermore, the paper studies to what extent WTP to improve supply reliability depends on the customers' perceived past experience and, more precisely, on the importance they give to the last outage they experienced in their homes. Finally, it investigates whether the WTP for renewable energy is related to energy-saving habits or to the level of concern about the GHG emissions stated by households. To achieve the objectives proposed in the previous paragraph, a stated preference choice experiment (CE) was conducted. In the CE, customers were asked to choose between two hypothetical companies that differed in terms of their attributes, or they had the option to stay with their current supplier. To reflect real-world choice situations, the current supplier is taken as a reference alternative, so hypothetical scenarios consisted of an upgrade in some attributes and/or a downgrade in other attributes as, for instance, in Banfi et al. (2008) or Longo et al. (2008). Nonetheless, and differing from previous studies in the electricity sector, labels are also used to characterize each of the alternative hypothetical suppliers rather than generic titles in order to increase experiment realism (Louviere et al., 2000). In addition, though the vast majority of past research has been based on fractional factorial designs, here the choice scenarios were generated using an efficient design (Huber and Zwerina, 1996), a novel technique which has been rarely applied in this area. To the authors' knowledge, this is the first time that an experimental design combining a reference alternative, labeled alternatives and efficient design is employed to analyze energy demand. The econometric model used is a mixed logit panel with error components with a focus on capturing systematic heterogeneity in the preferences. This model takes into account the panel correlation inherent in the data, the correlation between the experiment's hypothetical alternatives and the presence of random heterogeneity in the preferences. The effect of some specific factors analyzed in this work (e.g. environmental concern) is considered by introducing interactive covariates with rating-scale variables into the specification for the deterministic component of utility. The results help to predict the willingness of consumers to switch supplier in a market where few have exercised this option. This information might be useful for similar emerging markets and to compare with those territories where competition is at a more advanced stage. Furthermore, the results provide the first WTP estimates based on choice experiments in the context of the Spanish energy sector and highlight the importance of accounting for customers' perceived past experience, their electricity consumption patterns and their awareness of environmental issues when attempting to more accurately characterize the distribution of the population's WTP. The rest of the paper is structured as follows. In the next section, we describe the experiment of stated preferences. In Section 3, we present the microeconomic bases and the econometric formulations, while in Section 4 we provide a brief, descriptive analysis of the data used. Section 5 shows the results of the different models estimated and discusses the results. In Section 6, we present the estimated willingness to pay for the attributes analyzed in the context of existing empirical evidence for other markets. Finally, we offer the main conclusions drawn from this study.

Deregulation and the introduction of competition in the electricity industry in many countries have led companies to persistently pursue customer satisfaction in the retail market. The knowledge of customers' preferences and WTP is key information for suppliers in order to analyze the retail market development, and for governments, to design energy policies. In this study, we analyze these issues in an emerging market situation. Specifically, a stated preference electricity company choice experiment was conducted in the Canary Islands' residential market, an emerging electricity market where the environment plays a key role in the economic activity. As a novelty in the research area on energy demand, the experimental design combines a reference alternative, labeled alternatives and efficient design. The study provides valuable information on the likelihood of switching supplier in a market where few have exercised this option. Likewise, unlike previous studies on this subject, preferences for the electricity supplier and WTP for service attributes are examined considering additional factors, such as perceived past experience with the current supplier, perceived importance of last outage experienced at home, energy-saving habits or the level of concern about the GHG emissions stated by households. The results could be very useful for electricity systems with similar geographic and market conditions. The results obtained on supplier choice allow us to draw three conclusions: • Ceteris paribus, 50.7% of respondents prefer their current supplier, 40.1% prefer another supplier and only 9.2% prefer a company from another industry. Thus, new companies could enter the retail market even without improving significantly the level of service provided by the current supplier. • However, the majority of individuals who rate the service provided by current supplier highly would not change supplier (63%), and also a significant portion of those who rate a competing company highly (44%) would stay with their current supplier. Therefore, there is evidence of significant switching costs and/or brand loyalty to the current supplier. • The effect of brand loyalty on the current company and/or the perceived switching cost increase with the age of the customer. With respect to the WTP obtained, we highlight the following findings: • WTPs for all of the attributes increase with household income and their values are in the line of those obtained in the literature, which indicate that our specific study context does not affect these values. • Customers place a high value on having fewer outages and on these to be of short duration. • On average, individuals are willing to pay for an energy audit service with shared savings, although this would only be a small amount representing barely 2% of their monthly electricity bill. • Respondents with university studies are also more willing to pay for renewable energy. Specifically, regarding the new factors analyzed and their influence on the WTP for service attributes, the main results lead us to conclude that: • The perceived past experience of customers significantly affects the WTP for reducing power outages. We found evidence of a systematic variation in the preferences for outage frequency, in that the doubling of the WTP could be explained by respondents' experience in terms of the severity of power outages that they state having suffered in the past. • Customers who state having an active energy-saving behavior are those who attach a higher value to the introduction of renewable energy in the generation of electricity. In fact, these individuals exhibit a WTP that is up to four times higher than those who do not save energy. Also, a greater WTP for renewable energy was clearly evidenced in individuals who claimed to be concerned with GHG. These results should be interpreted as an indication of different aspects new firms need to consider if they enter the market. Specifically, it is possible to compete by offering services connected with the different attributes studied. However, they would have to deal with customer loyalty to the current supplier, as well as customers' costs to switch suppliers. Our results on the WTP are in line with those obtained in other studies conducted in other contexts and with different methodologies. One important fact is that the aggregated WTP for renewable energy would cover less than 50% of the expenditure required to fund the goals for these energy sources listed in the energy plan currently in effect. Therefore, the regional objectives in the use of renewable electricity generation cannot be achieved by using only market mechanisms as found by Bollino (2009), Batley et al. (2001) and Scarpa and Willis (2010) and Ek (2005). Furthermore, as it has been shown, the concern for GHG and energy-saving action behavior are key factors in influencing the WTP for renewable energy; thus, all policies that affect these two factors would lead to greater financing to achieve greater penetration of renewable energy throughout the market. For example, information campaigns about rational energy use and promoting awareness of environmental problems in the generation of electricity. Finally, other avenues for further research can be identified. It would be useful to provide more empirical evidence on these issues by replicating this study in other electricity systems where geographic and market conditions allow us to make accurate comparisons. Thus, we could test whether the special circumstances arising from the Canarian economy and geographical location have a decisive effect on the WTP for service attributes and on choice of supplier. Another extension includes a more accurate assessment of switching costs by studying a group of customers who have changed suppliers. Future research should also consider both groups of users, those who have switched and those who have not. This would also confirm whether an inertia effect exists involving previous suppliers, as has been shown in other sectors.