Economic regulation of electricity grids in Nordic countries

Transcription

1 Economic regulation of electricity grids in Nordic countries Report 7/2011

2

3 Economic regulation of electricity grids in Nordic countries Nordic Energy Regulators 2011

4 Report 7/2011 NordREG c/o Danish Energy Regulatory Authority Nyropsgade 30 DK-1780 Copenhagen V Denmark Telephone: (+45) Telefax: (+45) [email protected] Internet: December

5 Table of contents 1 Introduction Distribution of electricity is a natural monopoly Economic regulation The regulatory assignment Methods of economic regulation Some history of economic regulation in the Nordic countries Structure of distribution and transmission Denmark Finland Norway Sweden Grid structure in Nordic countries Economic regulation Goal and purpose the rules of economic regulation Denmark Finland Norway Sweden Differences and similarities in the electricity acts Regulatory period Denmark Finland Norway Sweden Structure of the regulatory model Denmark Finland Norway Sweden

6 2.4 Capital costs Denmark Finland Norway Sweden Overview of rate-of-return parameters and the RAB Operating costs Denmark Finland Norway Sweden Differences and similarities Quality regulation Denmark Finland Norway Sweden Overview of quality regulation Differences between regulation of DSO and TSO Denmark Finland Norway Sweden Applied benchmarking Methods and models Denmark Finland Norway Sweden Different methods and use Evaluation of regulation and issues in future regulation

7 6.1 Finland Norway Sweden References Country report: Denmark Number of DSOs in Denmark General description of the Danish revenue cap regulation Benchmarking of the DSOs cost efficiency Measuring quality of supply Benchmarking quality of supply among Danish DSOs Regulation of distribution network tariffs Variation in distribution tariffs Development in distribution tariffs The regulatory asset base Allowed rate of return Violation of the allowed rate of return Allowed profits Network losses Regulation of the TSO Country report Finland The objectives of regulation Regulatory period The differences in regulation between TSO and DSOs Incentive to improve quality Innovation incentive system Incentive to improve efficiency The current regulation in Finland Structure of the regulation model Inflation Calculating the reasonable return Calculating the adjusted profit

8 9.4.1 Incentive to improve quality Innovation incentive system Other adjustments of profit and loss account Applied benchmarking method Incentive to improve efficiency Model specification Contextual (operating environment) variable The reasonable (allowed) total cost level (STOTEX) Evaluation of regulation and issues in future regulation Appendix References Country report: Norway The objective of the regulation Introduction Direct regulation Tariffs Regulation period Current regulation Grid structure The DSO model Total costs Operation and maintenance costs Inflation Capital costs Network losses Quality impact Regulation of the TSO The Cost norm The DEA models Regional and central grid The weight system

9 Geographical data The DEA results and calibrations Evaluation of the economic regulation and future plans References Country report: Sweden The objective of the regulation Introduction Framework for the economic regulation Regulatory period Differences between regulation of DSO and TSO Economic regulation Ex ante regulation with revenue caps Transition from ex post to ex ante Capital cost Operating cost Quality regulation Applied benchmarking methods Issues in future regulation References

10 Preface The intention with this report is to give a short overview of the design of the economic regulation in the Nordic countries. NordREG has worked as an organisation for collaboration between the Nordic countries regarding regulation of the electricity market since The main mission is in cooperation, we actively promote legal and institutional framework and conditions necessary for developing the Nordic and European electricity markets. During NordREG organized internal workshops on the subject of network regulation of DSO and TSO (distribution and transmission system operators). In 2011 the NordREG board decided to permanent the work on this subject and started a new working group network regulation WG dedicated to the issues of regulation. During 2011 year the group has worked with a report which provides an overview of the regulation of DSO and TSO in Denmark, Finland, Norway and Sweden. The report has focus on the applying methods of economic regulation in the Nordic countries and the regulators tool-kits to calculate the revenue cap for each electricity company. For details on used parameters, sub reports, manuals, etc. the reader is recommended to visit each regulator s website. The working group has consisted of delegates from each regulatory agency: Herlita Bobadilla Robles, chair of the group (EI, Energy Markets Inspectorate - Sweden), Göran EK (EI), Matti Ilonen (EMV, Energy Market Authority Finland), Nicolaj Mølgaard Jakobsen (DERA, Danish Energy Regulatory Authority), Silje Cathrine Syvertsen (NVE, Norwegian Water Resources and Energy Directorate), Simo Nurmi (EMV), Siri Hau Steinnes (NVE). Finn Dehlbæk Chair of NordREG 8

11 Summary This report is about the design of economic regulation of electricity companies in the Nordic countries. The purpose is to inform the interested reader on how the regulation of tariffs is designed. The intention is to give a short overview on the current economic regulation with an ambition to focus on differences and similarities. A common feature of the electricity distribution sector is that the industry structure consists of many independent companies with great differences in size and density of customers. This is contrary to what is common in other countries. The regulatory task can be more challenging with many separate utilities to regulate, especially if the industry is very heterogenous. In the appendices, the economic regulation of each country is presented in more detail. In the main text the focus is on differences and similarities. When comparing the regulations one can make two observations. On a superior level there are great similarities. All Nordic countries regulate the network companies by setting revenue caps. The legislation, the goals given to the regulators and the regulators general interpretation of the rules are to a great extent the same. The primary purposes are to prevent the monopolist to overcharge customers and to create a rational network industry. The regulation shall stimulate an effective management resulting in productivity development and optimal quality of the services. The differences in the Nordic economic regulations are in the details in the setup of the regulatory models and choice of parameters. For instance; the assessment of a reasonable rate-of-return is done in all the regulations. When deciding this rate-of-return some countries use the method of weighted cost of capital (WACC) and capital asset pricing model (CAPM), other do not. Even when using the same method, the inputs in the model are not the same. Common for the regulatory models are the division of costs related to capital costs and operating costs. The latter costs are in turn divided in controllable and non-controllable operating costs. The assessment of the regulatory asset base differs between the countries and the different methods imply rather big differences in the calculation of revenue caps. How to include network losses is also treated somewhat different between the countries, either as a controllable or non-controllable cost. The use of benchmarking and assessment of efficiency targets also varies between the countries. Some countries make use of specified benchmarking models, while others use a general efficiency target. The efficiency target or the result of a benchmark process is company specific in Denmark and Norway, in Finland there are both a company specific and a general target (same for all DSO) and in Sweden there is a general target. All countries have included quality of supply in the economic regulation, but the impact of quality of supply in the economic regulation differs. Common for all four countries is the treatment of very long interruptions. Direct schemes for compensating those customers are in use. For the future regulation the regulators are working with development of the tools for the quality regulation. More precise data collection on outages and better estimates on the valuations of outages for different group of customers will be introduced in the future. 9

12 10

13 1 Introduction 1.1 Distribution of electricity is a natural monopoly In some industries, like the electricity network industry, it is inevitable that market power arises. A prominent reason is that there are increasing returns to scale in infrastructure; so large that duplication of the infrastructure is economically impossible or at least costly enough to make an infrastructural monopoly preferable. In the economics literature, industries of this kind are known as natural monopolies. Economic regulation is used for controlling the existing market power in particular industries. Economic regulation typically targets the prices of the product or services. This can be done either by directly regulating the consumer prices for the final product or by regulating the prices for access to the infrastructure in the case of third-party-access. Alternatively, the pricing can be indirectly regulated, e.g., by capping the company s profit level or its return on capital. 1.2 Economic regulation There are several methods of economic regulation of natural monopolies. Today most sector-specific regulations apply ex ante. The EU directives do not permit an ex post regulation. 1 There has been a move from ex post regulation of primarily public owned infrastructural utilities to ex ante regulation. An ex ante regulation means that the method for the economic regulation of prices or revenues is decided by the regulator and known in the regulated industry before the regulatory period begins. Ex ante regulation is usually not completely fixed. Some parameters, for instance the rate of return, prices on power, inflation etc., can be updated during the regulatory period. An ex ante regulation can be designed in different ways. The regulation can specify each individual price, it can apply to an average of prices or it can cap the regulated companies rate of return on assets or the total revenues. The regulation is aiming at controlling variables like prices or revenues. To do that, the regulator needs knowledge on productions costs and the objective factors for production for each monopoly. The challenge of economic regulation is mainly that the regulator does not know the true level of efficient production and costs for each company, that is; there is asymmetry of information. Other variables like quality of the output also have an important impact on the economic regulation The regulatory assignment The task of economic regulation is to balance the objectives of both grid owners and customers. The owner wants a certain minimum return on the assets and the customers want low tariffs. The regulators task is, in terms of economic theory, to maximise the welfare for society. That is to design the regulation which will maximise the sum of the customers and producer surplus. The regulator has to decide on several parameters in 1 Article 37 1a, Directive 2009/72/EC of the European Parliament and of the Council of 13 July 2009 concerning common rules for the internal market in electricity and repealing Directive 2003/54/EC. 11

14 designing the regulation. The parameters have different impacts in the short and long run. For instance will a low rate of return benefits the customer in the short run, but can eventually be problematic in the long run if investments decrease due to the low rate of return Methods of economic regulation Several methods of regulation of infrastructural services have been used and are in use: rate of return, cost-of-service, price cap, revenue cap, yardstick regulation, performance standards, earnings- sharing (sliding scale) etc. The different methods have pros and cons and usually a regulation is designed with features from different approaches. It is common to use combinations of these, for example a revenue cap regulation with inflation index and yardstick analysis in combination with bottom (minimum) and ceiling (maximum) rules on rate of return. The difference between the methods can be viewed in terms of which part takes the risk of future changes. In a pure rate of return case the risk falls heavily on the customer, while in a price cap regulation the risk falls more on the producer. But a price cap case also gives incentives to be more productive, since all benefits from rationalisations with a given price as the restriction will be a reward to the producer. The approach with price or revenue caps was applied as a way to stimulate the companies to be more efficient. Therefore the term incentive regulation was set for regulatory methods with focus on cost efficiency. The development of economic regulation goes from cost-of-services and rate-of-return regulation to different variants of capping methods. The capping can be fixed in prices or revenues. For electricity grids the capping is generally done in revenues. The capping can also be combined with benchmarking where the outcome in cost and production are compared to some norm level. Regulation with a benchmark element is called performance based regulation or yardstick competition (yardstick regulation) Some history of economic regulation in the Nordic countries The deregulation in the Nordic countries started in Norway in 1991 followed by Finland in 1995 and Denmark and Sweden in There are several similarities between the countries, like the design of the electricity market and the separation of the competitive and natural monopolies parts of the electricity industry. Regarding the design of the economic regulation of the distribution and transmission network there are rather big differences. Brief history of economic regulation in Denmark The Danish deregulation was implemented in The regulation was at that time a rate-of-return regulation. Since January 2005 is the economic regulation designed as a revenue cap regulation. Brief history of economic regulation in Finland Associated with the electricity market restructuring in 1995, the basis for the electricity distribution regulation was formed. However, no exact regulation methodologies were defined. The Finnish Electricity Market Act (386/1995) only stated that distribution pricing shall be reasonable and distribution companies shall operate efficiently. 12

15 The regulator evaluated the reasonableness of distribution pricing ex post on the basis of the observed rate-of-return. During the first investigation case on the reasonableness of distribution pricing, the basis of the regulatory model was formed. The first decision was legally effective in 2000 (so called Megavoima-decision). Only the companies that were suspected of overcharging customers were investigated. The regulator had no formal legislative power, and its decisions were always case-specific. The EU legislation defined also network regulation in Finland. The first regulatory period beginning in 2005 started the ex ante electricity distribution regulation in Finland. Brief history of economic regulation in Norway The Energy Act came into force in In 1993 a rate-of-return regulation was introduced as a temporary model while an improved regulation was developed. A revenue-cap regulation was implemented in 1997 and has been the foundation of the economic regulation in Norway till today. A general efficiency requirement was introduced with the revenue-cap model, and an individual efficiency requirement based on benchmarking analyses (DEA) was implemented in 1998 for the distribution grid and in 1999 for the regional grid. Incentives for quality of supply were introduced in the economic regulation in In 2007, the revenue-cap model in use today was introduced. Brief history of economic regulation in Sweden The electric power distribution in Sweden has been subject to ex-post regulation up to From 2012 there will be an ex ante regulation. The deregulation in Sweden started in 1996 with the separation of production and retail from transmission and distribution. The first regulatory model for DSO/TSO was built on the principles of cost-of-service. The principle of cost-of-service regulation for public utilities in the municipality law was the point of departure. 3 The majority of the DSOs were owned by municipalities. The cost-of-service regulation in combination with public ownership constituted a rather light handed regulation. This form of light handed regulation was of ex post type. The companies set the tariffs and the regulator could on its own initiative or by complaints from customer start a case. The regulatory legislation introduced in 1996 the regulator took the price level of the tariffs from that year as a reference. In 2005 the Swedish regulator implemented an incentive regulatory scheme for the DSO by which actual revenues would be benchmarked against the cost of an engineering designed cost norm model. As a consequence of the critique from the European Commission regarding the lack of legal certainty in the use of the NPAM and the appeal to the administrative court the NPAM regulatory model was dismissed in A commission worked on the issue and proposed a new legislation for an ex ante regulation. 4 After the parliamentary decision on 2 The implementation of the net performance assessment model in Sweden intended to fulfil the EU directive. But this way to decide on a calculation model in advance, but to do the very regulation ex post was not permitted by the commission. 3 Damsgaard N and Green R, Den nya elmarknaden framgång eller misslyckade?, SNS förslag SOU 2007:99, Förhandsprövning av nättariffer m m,

16 the new legal framework the regulator developed a new model with an ex ante feature that will come into force in January Structure of distribution and transmission The industry structure of the distribution grid in the Nordic countries has great similarities contrary to several other countries in Europe. 5 The distribution industry is characterized of many independent grid companies with big differences in both size and density of customer. The structure can be described by a Salter graph, which is shown for each country below. The graph has customer density on the vertical axis measured as meter per customer and size of company in terms of number of customers on the horizontal axis. The transmission grid industry has one company in each of the Nordic countries. They are either fully owned by the state or the state own the majority. There also exists an intermediate level between transmission and distribution level which does not exist in the other European countries Denmark Denmark has one transmission system operator called Energinet.dk, eight (8) regional subtransmission system operators ( kv) and 84 distribution system operators (0,4 20 kv) in Denmark. There is a significant variation in size, ownership-structure etc. among the Danish DSOs. The largest DSO DONG Energy Distribution has more than a million customers which is more than 20 %age of the Danish population while the smallest DSOs only have a couple of thousand customers. There have been relatively many mergers among the Danish DSOs during the last couple of years and 15 DSOs are relatively large compared to the remaining Danish DSOs. In Figure 1 is the industry structure illustrated. 5 With the exception of Germany with about 700 independent DSOs, the majority of the DSO industries in Europe have of a few DSO (1-15 DSO). 14

17 Figure 1 Structure of the Danish distribution: customer density and size Finland At the end of 2010 there were 87 distribution grid operators (DSO), 12 high-voltage (regional) distribution grid operators and one transmission system operator (TSO) in Finland. The TSO is Fingrid Oy. The mean value for number of DSOs customers was just over The mean value for the length of the distribution grid was 4370 km. The variation between grid operators is significant. The smallest DSO by customer base is serving only 22 customers and the largest has customers. The company with the least customers is excluded from the Figure 2 below because of the very low density (1650 metres per customer). Two other DSOs are also excluded due to low density (424 and 403 metres per customer respectively). 15

18 Figure 2 Structure of distribution operators in Finland: customer density and size The mean value for density is 169 metres per customer. The DSO with highest density is Helsinki Energy with 18 metres per customer. The DSOs with largest market share are Fortum with a market share of 19 % and Vattenfall with a market share of 12 %. The customer density for these two DSOs is metre per customer. The mean market share is quite low (1,2 %) and there are only three DSOs with a market share of over 10 %. The three largest DSOs have together a market share of 41 %. The DSO with a low density is typically small which can be seen in the Figure 2 where the very thin bars are more frequent in the right part of the graph. The grids operated by Finnish DSOs consist of 0,4kV, 1-70kV and 110kV lines. About 80% of the 110 kv grids in Finland is operated by DSOs and the rest is operated by the high voltage (regional) distribution grid operators Norway The electricity grid in Norway is divided into three levels similar to the other Nordic countries: distribution grid, regional grid and central grid. The distribution grid is defined as the networks with voltages up to and including 22 kv. The regional grid is a sub transmission system between the distribution and central grid, with voltages from 33 kv up to and including 132 kv. The central grid includes all voltages above 132kV, still mainly a voltage of 300kV and 420 kv, but also some 132 kv lines are included. The central grid is a national transmission system. The transmission system operator (TSO) is Statnett SF. The TSO owns most of the central grid (approximately 90 % in 2011). Further, there are 158 network companies (in 2009), whereas 142 have distribution grid and 91 have regional grid. Several companies have both distribution and regional grid, and a few also own some of the central grid. The largest company in the distribution grid is Hafslund Nett AS, with 541 thousand customers, GWh energy supplied and km lines defined as distribution network. There are several very small grid companies. In most of the cases their main 16

19 activity is production of electricity or some kind of industry, but in addition they own some network that is obliged to be regulated. One of the smallest companies that can be considered as a real distribution company is Modalen Kraftlag BA with 387 customers, MWh energy supplied and 68 km of distribution lines. The structure is shown in Figure 3. Figure 3 Structure of distribution grids in Norway: customer density and size The graph shows the customer density of most of the DSO companies in Norway. About ten of the smallest companies explained above are excluded. They have an extremely high or low density due to nearly no customers or no lines. The mean value of density is 299 meter per customer if all companies are included, but 157 meters per customer if the 10 extreme companies are excluded as in the graph. If we look at the latter sample, there is a variation from 43 to almost 280 meter per customer The largest DSO Hafslund Nett AS with distribution in the Oslo area, has a market share of 20 %. It is also the only DSO with a market share above 10 %. The mean market share is only 0,8 %. Of 129 DSOs, 20 utilities have less than 2000 customers Sweden The Swedish electricity grid is divided into three levels depending on the voltage level: transmission, subtransmission (or regional) and distributional (local level). The transmission level is operated by one state owned company (Svenska Kraftnät) with lines with a voltage of 220 or 400 kv. The subtransmission level (regional grids) is the link between transmission and distribution. The electricity is transformed to a lower voltage level, usually kv. The subtransmission level consists of four companies (the three biggest distribution companies in the country and one small in northern Sweden (Skellefteå). 17

20 Local grids are connected to regional grids and typically operate with a voltage level of 0,4-20 kv. The distribution to consumers goes normally via the local grid with the exception of a number of major industries with high levels of consumption of electricity, which are connected directly to the subtransmission grid. The statistics in Figure 4 is from Customer density varies from 22 to 359 metres per customer. Mean value is 110 metres per customer. Size varies from 18 to customers. Of the 20 companies with fewer than 2000 customers the density varied from 39 to 359 metres per customer. The development goes towards fewer and larger companies. At the end of the 1950s, there were more than 1500 companies and in the early 1980s the number had dropped to 380 companies. The two main areas are Vattenfall Eldistribution AB (REL00583) with customers and E.ON electricity networks Sweden AB (REL00593) with customers. Customer density of Vattenfall was on average of 109 metres per customer and for the E.ON area 122 meters per customer. Because it is a very large area with both urban and rural areas the customer density will vary a lot within the distribution area. The E.ON and Vattenfall groups have several DSO areas. The market share for the three dominant companies is 51,4 % (E.ON 19,0 %, Fortum and Vattenfall 16,2 % each). The mean value for market share for the year 2009 is 0,6 %. Figure 4 Structure of the Swedish distribution: customer density and size In Figure 4 are two DSO with the lowest density excluded Grid structure in Nordic countries In Figure 5 the customer densities in the Nordic countries are put together in the same graph, and shows from the left Denmark, Finland, Norway and Sweden The mean value of density is 51 meters per customer in Denmark, 169 meters per customer in Finland, 18

21 157 meters per customer in Norway, 110 meters per customer in Sweden. The proportion of small grids (less than 2000 customers) is on the other hand higher in Norway than in the other countries. It has to be mentioned that the DSOs in Finland own grid with higher voltage levels compared to the other countries and this probably gives a picture of less density in Finland. There are very few costumers connected to high voltages lines. In the data from Finland lines for 0,4 kv up to 110 kv is included. About 80 % of the 110 kv lines in Finland are operated by the DSO companies. The graph for Finland is to some extent skewed. The density of distribution seems to be more rural in Finland than the other countries. The statistics for the other countries are easier to compare because they only have lines in the voltage levels of 0,4 kv to 20 kv in Denmark, 0,2 kv to 22 kv in Norway and 0,4 kv to 20 kv in Sweden. A difference in Norway compared to Finland and Sweden is that in Norway the big companies mainly operates in urban areas, whereas in Finland and Sweden the big companies like E.ON, Fortum and Vattenfall operate in both rural and urban areas, and with a relative high proportion of rural areas. It is still Sweden that has most DSO companies with 175 utilities, while Norway has 142, Finland 87 and Denmark 84. The mean value for market share is low in all countries; 0,6 % in Sweden, 0,8 % in Norway and Finland with 1,2 %. The existence of a few very big companies gives another picture of the concentration of the industry. The three biggest companies have a market share of 41 % in Finland, 33 % in Norway and 51 % in Sweden. The regulators thus have a complex regulatory task when it comes to regulate a few very big companies and many small companies. The distribution services are also produced in different environments regarding geography and climate. 19

22 20 Figure 5 Grid structure in Nordic countries regarding density and size

23 2 Economic regulation The design of the countries regulation is described in this chapter. A common feature for the distribution sector is that the industry consists of many independent companies contrary to what is common in other countries. The regulatory task will be different with many separate agency s to regulate. When the industry also is heterogenous with significant differences regarding size and density of customer the regulatory task becomes even more complicated. 2.1 Goal and purpose the rules of economic regulation In this chapter we describe and compare the legislative structure in the Nordic countries - the rules for the economic regulation. There are several levels for the rules of regulation: The EU-level with the directives, rules given by the parliament in each country and interpretation and application of these rules by the regulator. 6 The purpose and goals of the regulation are to a great extent the same. The primary purpose is to protect the customers for being overcharged and to create a rational industry. The regulation shall stimulate an effective management of the company that will lead to productivity development. The goal is to balance the goals of the companies and customers wish for low tariffs and high quality of supply Denmark The mission is to protect the customers for being abused and to work for efficiency in the distribution and transmission parts of the electricity market Finland According to EU legislation (Directive 2003/54/EC), national regulatory authorities are to ensure that transmission and distribution tariffs are non-discriminatory and cost-reflective. In addition, the directive further states, that the distribution system operator shall maintain a secure, reliable and efficient electricity distribution system in its area with due regard for the environment. And finally, distribution tariffs should be sufficient to allow the necessary investments in the networks to be carried out in a manner allowing these investments to ensure the viability of the networks. According to Section 38 a 1 of the Electricity Market Act, the Energy Market Authority shall confirm for every network operator separately the methods assessing the return on network services and transmission service charges during the regulatory period. The methods shall be confirmed before the implementation (confirmation decision). According to Section 38 a 2 of the Electricity Market Act, confirmation decision shall be based on the principles prescribed in Section 3, 4 and 6 of the Electricity Market Act and Act of the European Parliament and European Council (EC No 714/2009). The regulatory methods concerning the pricing of electricity distribution operations and high- 6 Norway is not a member of EU but they follow the directives due to EEA agreement 21

24 voltage electricity distribution operations are based on the pricing principles stated in Section 3 10 and 14 of the Electricity Market Act. According to Section 38 a 2 of the Electricity Market Act, in the confirmation decision Energy Market Authority can determine about: The valuation principles of the capital invested to network operations The methods of measuring reasonable return for the capital invested to network operations Methods of determining the result of the system operations and the correction of the income statement and balance sheet required by them Target encouraging improvement of the efficiency of the system operations and the method of determining it, as well as the method to apply the target in pricing Norway The regulatory frame for economic regulation of the grid companies is formulated based on objectives given in the Energy Act and regulations laid down in accordance with the act. The key objectives are according to the Energy Act. Regulation, Section 4 4-1: ( ) to secure a socially rational energy sector and network through enabling an effective energy market and an effective management, utilization and development of the electricity network ( ) The main principles for the economic regulation are further defined through the Energy Act Regulation, where it is stated that revenue caps shall be determined annually and that the income over time shall cover network costs and give a reasonable rate of return on invested capital provided that the network is managed, utilized and developed effectively. In NVEs underlying regulation it is further stated that the purpose of the economic regulation is to secure that power is transmitted at correct quality of supply and price, and that the network is utilized and developed in a secure and socially rational manner. Economic and direct instruments must be combined in the regulation of power networks in order to achieve an optimal behaviour of the utilities and at the same time avoid unwanted external effects. This is why the power networks are subject to many sets of rules in laws, regulations and conditions for licensing which regulates their duties and rights. These regulations leads the power networks activity, and shall secure that the companies invest and maintain the network, that safety- and emergency preparedness concerns are ensured, that the quality of supply is sustained adequately, that the security of supply is maintained in demanding situations, that the electrical constructions are adequately robust, and that investments with large environmental disadvantages are not carried through if the society s benefit from the investments are lower than the social costs. NVE expects that the network companies follow the direct regulations. The compliance of the direct regulations is controlled through a comprehensive monitoring activity each year. 22

25 2.1.4 Sweden Grid companies are regulated in addition to the usual rules of company law, tax rules also by special legislation on grid activity contained in the electricity Act (1997: 857). The new ex ante regulation beginning in 2012 has focus on the revenues. The regulator shall decide on each grid companies revenue caps after a proposal from each company. The revenue cap shall cover reasonable operational costs and a reasonable return on the assets used in the distribution/transmission. Quality norms are integrated in the cap so if norm values for delivery (outages) are exceeded (lowered) during the regulatory period, some reductions (rewards) in the next revenue cap will be the decided. The purpose is to give incentives for future improvement in quality. According to Section 5, 1 of Electricity Act the revenues will be fixed in advance for each regulatory period consisting of four calendar years, unless there are special reasons to use another period of time (4 ). In the decision of the revenue cap the data and methodology used in the determining the revenue cap should be described (3 ). A decision applies without prejudice to an appeal. The revenue cap states that the cap should cover the reasonable costs of conducting grid activities during the supervisory period and provide a reasonable return on capital (equity) needed to carry out the activity (6 ). Regarding the design of the tariffs the legislation says that: Grid tariffs should be objective and non-discriminatory (Section 4, 1 law 2009:892). There is no embracing paragraph on the purpose of the regulation in the legislation or in governmental decree to (EI) Differences and similarities in the electricity acts The purpose of the regulation and the regulators are to a great extent the same in the Nordic countries. The regulator shall comply with the given legislation and from the legislative rules, design a regulatory model to be applied on the network industry. The differences in the legislations are in the details. For instance, in Denmark a general efficiency target is prohibited, while in the other countries the law does not address this issue. 2.2 Regulatory period In the Nordic countries, the regulation tends to work for a period of 4-5 years, except Denmark that only has a period of one year. Common for the design of regulation is the ex ante decision on the principles (method) for the revenues. The distinction is how the regulator update the data and revenue caps during the regulatory period; some makes new calculation every year and sets yearly revenue caps and others sets all parameters in the beginning of the regulatory period and adjust these only at the end of the period Denmark The regulatory period is one year. Energitilsynet shall each year decide on a yearly revenue cap for all grid levels (TSO, subtransmission and DSO) Finland The length of regulatory period is four years. But the first period with ex ante regulation was limited to three years. First period was The second was and the third will be

26 2.2.3 Norway The economic regulation of the grid companies is executed by annual determination of revenue caps for each company. The current revenue cap regulation entered into force as of 1 January According to the Energy Act Regulation, the main principles for calculation of revenue cap shall be re-evaluated periodically. Each period shall last a minimum of 5 years. However, minor model adjustments can be carried out continuously over the period. The regulatory model also implies that some parameters are updated yearly the annual revenue cap takes into account yearly changes in CPI, interest rate on government bonds used in determination of the WACC and in power prices used to calculate costs related to network losses Sweden From January 2012 there will be a change to an ex ante regulation. An approval of the grid companies proposal of revenue caps will take place in advance. The regulatory period is four years and the first regulatory period is For the TSO the regulatory period is one year. Up to 2012 the regulation has been done ex post. So each year has been treated as a regulatory period. 2.3 Structure of the regulatory model To get an overview of the countries regulatory models some graphics are presented Denmark The regulation in Denmark is designed as a revenue cap with a maximum rate of return on network assets. If a DSO does not violate its revenue cap and the maximum rate of return on network assets, the DSO is free to set its distribution tariffs as long as the DSO does not discriminate between different types of customers. Some DSOs choose not to fully exploit their revenue caps. Among other things, this has created a large variation in distribution tariffs among the DSOs. For each DSO, the necessary costs are defined as the level of operational costs, depreciations and capital costs associated with an efficient operation of the distribution network in Each year, a DSO s revenue cap is basically determined by the DSO s price-adjusted necessary costs in A DSO s revenues cannot exceed the DSO s price adjusted necessary costs from 2004 as a general rule. A DSO s revenue cap is permanently increased with the increased costs of network losses, while the DSO s revenue cap is only temporarily increased with some allowances due to a necessary investment as shown in Figure 6 below. 24

27 Figure 6 Overview of the regulation Finland The Finnish electricity distribution regulatory model can be described as an ex-ante revenue cap model. EMV uses the regulatory model to set the reasonable rate of return and the DSOs set the tariffs themselves. The regulatory model is built of many components and Figure7 illustrates the formation of reasonable return and actual adjusted profit. The reasonable rate of return is subtracted from the actual adjusted profit annually and the result is surplus (+) or deficit (-). After the regulatory period, the 4-year total surplus/deficit is calculated by adding up the yearly surpluses/deficits. DSOs are obligated to compensate the surplus and allowed to compensate the deficit in their price setting on the following regulatory period. EMV uses both building cost index and consumer price index for indexation in the regulatory model. 25

28 Figure 7 Structure of regulatory model of DSO in the third regulatory period Norway NVE regulates the network companies by setting an annual revenue cap. The regulation can be described as incentive based, where 40% of the companies actual costs (cost base) are more or less passed through while 60% are based on cost norm where benchmarking analysis are used to describe the companies cost efficiency. If a company is defined as efficient due to the benchmark; 100 % of the costs are covered. The cost base that enters the revenue cap is based on data two years back. NVE notifies the estimated revenue caps before the year starts, based on estimates for WACC, CPI and price on power. When the year is over, NVE sets the final revenue caps based on observed parameters. In addition, property tax and tariffs costs to other grids are added to the revenue cap, actual CENS costs are subtracted and all changes in capital costs in the revenue cap year are added or subtracted when setting the allowed revenue. The allowed revenue constitutes the basis for each company s calculation of tariffs and the revenue compliance is subject to regulatory control. The excess/deficit between the actual collected revenues and the allowed revenue is to be adjusted towards zero over time. The regulation makes sure that a company measured as average efficient by our benchmark analyses earns a return on capital equal to the WACC. This also regards the industry as a whole, the cost norm are calibrated to make sure that the total revenue cap for the industry equals their total cost. The revenue caps are based on costs two years back which are price adjusted to reflect estimated costs in the revenue cap year. When the actual costs in the revenue cap year are known, the deviation between estimated and 26

29 actual costs are added/subtracted to the revenue cap two years later to make sure a return on capital equal to the WACC. The structure for the calculation of revenue cap and allowed revenue is shown in Figure 8. Figure 8 Calculation of revenue cap and allowed revenue Sweden The regulatory model of Sweden is structured on the different cost items. First the division between capital cost (CAPEX) and operating cost (OPEX). Then latter cost is in turn divided in controllable and uncontrollable cost. The controllable cost is reduced year by year by an efficiency target (a general requirement on lower cost for given output). This requirement on higher productivity is not applied for the uncontrollable cost. For capital cost the assessment of the regulatory asset base is the first and important part. The regulatory asset base is valued by the principle of replacement value. The norm for rate-of-return is decided by the method of WACC. The quality is set as a norm for the period in form of a mean value for historic data on interruption (SAIDI and SAIFI). The outcome of quality after the regulatory period is compared to the historically norm and the return on capital is adjusted in relation to the change of quality. The structure is shown in Figure 9. 27

30 Figure 9 Structure of regulatory model for Sweden 2.4 Capital costs The capital costs depend on the value of the regulatory asset base and the regulatory rate of return on capital. There are differences between the countries in how the regulatory asset base is constructed. To calculate the return on the regulatory asset base, a common approach is to use the method of weighted average cost of capital (WACC) where the equity part is calculated by the capital asset pricing model (CAPM) Denmark Asset base The initial regulatory asset base was determined in For a large number of grid components, DERA estimated unit investment costs and life-expectancy for each of the grid components. The initial regulatory asset base for a DSO was then calculated by multiplying the DSO s stock of each grid component with the associated estimated unit investment costs adjusted for depreciation. Each grid component is depreciated straight line. Investments that have been made after the year 2000 are added to the initial regulatory assets base with the actual investment cost. Thus, the unit investment costs were only applied to value grid components installed before year The estimated lifeexpectancies for the different types of grid components are, however, applied to straight line depreciate both the DSOs assets installed before year 2000 and from 2000 and onwards. The regulatory asset base for a DSO in 2001 is thus equal to the initial regulatory asset base from 2000 minus depreciations plus the actual costs of new investments made in The regulatory assets base for a DSO in 2002 is then equal to regulatory asset base from 2001 minus depreciations plus the actual costs of new investments made in 2002 and so forth. 28

31 Rate-of-return The DSOs revenues are furthermore regulated with a maximum return on grid assets which each year is determined by the interest rate on the 30-year building bond plus 1 %. Unlike in many other countries, we thus do not apply the Weighted Average Cost of Capital (WACC) to determine the allowed rate of return on capital. A DSO s annual result from operating its distribution grid divided by the DSO s regulatory asset base may not exceed the interest rate on the 30-year mortgage bond plus 1 %. According to DERA (2010), the interest rate on 30-year mortgage bonds plus 1 % was equal to 7,5 % in Accordingly, a DSO s annual result from operating its distribution grid divided by the DSO s regulatory asset base had to be less than 7,5 % in As previously explained, the regulatory asset base is determined annually. Since it is only the price-adjusted costs of capital and depreciations from 2004 that are included in a DSO s revenue cap, there is only a weak connection between a DSO s revenue cap on the one hand and the maximum return on grid assets and the DSO s regulatory asset base on the other hand Finland The Reasonable rate of return is calculated by multiplying regulatory asset base (RAB) with WACC - %. RAB consist of equity and interest bearing debts. Asset base The value of the network is calculated by its net present value (NPV) instead of its book value. The Energy Market Authority (EMV) does not apply the book value of the electricity network because this will not necessarily reflect the actual market value of the capital invested in the network. The NPV of the electricity network is determined each year, January first and this value is used in calculation of the value of capital invested in the network operations. When determining the capital invested in network operations, the electricity network administrated by the network operator is treated in the same way regardless of whether it is owned or leased by the network operator. When calculating the replacement value (RV) the unit prices of component-specific network components used in the value determination of the network are mainly based on the unit prices presented in the network recommendation of the Finnish Energy Industries (standard unit costs) and, as far as the components used in the above-mentioned value determination of the electricity network are not included in the above mentioned list of costs, the costs are based on the unit price study concerning the components in question, ordered from Empower Oy by EMV. Standard values are defined also for the buildings, sites, computer systems etc. invested in network operations. The standard unit prices are in the value of money of year 2010 and they are adjusted to correspond with the value of money of year 2012 (in the beginning of third regulatory period). NPV is calculated from RV using component-specific unit price and age data. The lifetimes of network components are based on lifetimes chosen by network operators in the first and second regulatory period. Network operator can choose a lifetime in the third regulatory period for those components that it has not chosen a lifetime in the first or second regulatory period. Lifetimes of network components vary between 5 and 60 years. The RV of the whole network is calculated by multiplying all the components with their respective unit prices. 29

32 There is a straight-line depreciation calculated from the replacement value of network (incentive to invest part of the regulatory model, see Figure 67). Inventories and account receivables (part that is committed to network operations) are appreciated on their book value in RAB. The amount corresponding to five % of the DSOs net sales is taken into account in the balance sheet under items Short and Long- Term Receivables, Financial Instruments and Cash and Bank Receivables as the necessary financial assets to safeguard network operations. The cost arising from these necessary assets is calculated by multiplying the above-mentioned share of the financial assets by the cost of interest-bearing debts used in the calculation of the WACC %. Rate of return Based on several expert opinions, EMV has selected the Weighted Average Cost of Capital model (WACC) as the basis for determining a reasonable return on capital invested in electricity network operations. EMV has applied the WACC model in assessing the reasonableness of pricing of electricity network operations since 1999, and the Supreme Administrative Court and the Market Court have confirmed with their decisions that the Authority s assessment method complies with the Electricity Market Act. The fact that the WACC model is also used as part of the supervisory methods in several other countries also speaks in favour of using the model. WACC parameters are shown in Table 1. Table 1 WACC parameters for the third regulatory period WACC parameter DSO TSO Risk free rate (real value) 10 year Finnish government bond yield (average of previous year May) - inflation component Inflation component 1,0 % 1,0 % (subtracted from nominal risk free rate) Asset beta 0,4 0,4 Equity beta 0,527 0,844 Market risk premium 5 % 5 % Premium for lack of 0,5 % 0,5 % liquidity Capital structure 30 / / 40 (Debt / Equity) Tax level 26 % 26 % Debt premium 1,0 % 1,0 % 10 year Finnish government bond yield (average of previous year May) - inflation component Norway The capital costs that enter into the regulated cost base are annual depreciations and return on the regulatory asset base (RAB). The RAB is based on historical book values 30

33 for the utilities fixed assets and other capital equipment used in the production of grid services. All accumulated depreciations and write-downs are subtracted. Fixed asset under construction are excluded from the RAB until it is activated. When calculating the return on capital, NVE use book values per 31/12. Asset base The RAB consists of the following fixed assets: transmission-, regional- and distribution grid components, sites, buildings, customer-specific equipment, means of transport, fixtures, tools and computer equipment, other fixed assets and allocated common assets. Leased assets are not included in the RAB, but in the operating costs. Assets that are financed by third parties as investment contributions, are not included in the RAB, however, operating costs related to these are not excluded from the overall operating costs. To reflect the working capital, NVE add 1% of the total book value per 31/12 to the RAB. The RAB is then multiplied with the WACC (described below) to calculate a reasonable return on the regulatory asset base. The depreciation of capital is linear. The companies decide the economic life of components based on their own experience, and then calculate the depreciation according to the expected technical lifetime of the specific grid component. Asset additions in the last and current year (year t and t-1) are not included in the RAB as a result of the time-lag in the cost base. However, the companies can calculate the revenue from investments into their allowed revenue the year they are commissioned. This implies that investments are also included in the companies tariff base as of the commissioning year. Rate of return The rate of return on capital is calculated using a Weighted Average Cost of Capital (WACC). The parameters are given in Table 2. The WACC formula is post tax, but NVE operates with a pre-tax rate of return. The WACC-formula is rewritten to an easier formula for pre-tax calculation of the rate of return: The formula is determined in the regulation on regulation of power grids. All parameters in the WACC formula except for the risk-free rate are fixed. Any amendments in the parameters require an amendment in the regulation. The risk-free rate on the other hand is updated annually when setting the revenue cap. The capital costs which consist of calculated return on the regulatory asset base and annual depreciation enter the revenue cap model as other costs where 40 % are passed though and 60 % are subject to a benchmark analyses. The regulation states that the companies shall achieve a reasonable return on capital, given effective management, utilization and development of the grid (see chapter ). An average efficient company earns a reasonable return on capital defined by the WACC or the NVE rate. It is decided that a company shall at minimum earn a return of 2 % over the last five years. Any company that falls below this minimum return will get a correction in its excess/deficit revenue balance to maintain at least a 2 % return on capital. 31

34 Table 2 Parameters for rate-of-return in Norway WACC parameters Risk free rate (nominal value) Annual risk free rate of government bonds with 5 years maturity Asset beta 0,35 Equity beta 0,875 Market risk premium 4 % Capital structure (Debt / Equity) 60/40 Tax rate 28 % Debt premium 0,75 % Sweden Asset base The capital cost is a product of a decided rate of return on assets and a capital base. The main principle is to use replacement values with real annuities. Assessment of capital cost starts with an estimation of a capital base. This value can be estimated by four methods where the replacement value by standard values for the different assets in use is the first choice. If there is no norm value specified for some kind of physical asset, the grid owner can use the estimations in the following order. Valuation according to norm values Valuation according to value at investment time Valuation according to book value Valuation by other way The basis for calculating the reasonable capital expenditures are the replacement value for the fixed assets the grid company use. The assets which can be included are specified in paragraph 3 of the regulation for determining the revenue cap according to the law. EI's report EI R2010: 07 discuss the valuation of electricity grid companies the different methods and how electricity grid companies should calculate and evaluate the capital. For more information, see also EI's PM 2010: 11 and 12, 2010: 2010: 13. Regardless of the method of valuation should therefore be calculated and replacement value plants are accounted for in the cost level for the year For more information about the accounting of capital base is set out in Chapter 5 in EIFS2010: 6 (regulation on the transmission of data). Rate of return Weighted Average Cost of Capital (WACC) is an established method for determining the rates-of-return in a business. WACC is used by the many regulators for decision on the norm value of rate of return. That is the reasonable return on capital that is both necessary 32

35 and enough for attracting capital without exploiting the customers. For the regulation WACC will be used as method for decision on the norm value for the rate of return. EI has asked two financial consultancies on their expertise to give an interval (minimum to maximum) for a reasonable interest. The discussion has been focused on how and how much the interest shall be influenced by the fact that the companies have untaxed reserves. EI commissioned two consultancies to do an estimation of a WACC for the regulatory period. The two reports were analysed and discussed and as a result there were some changes in the parameters. In Table 3 an overview of the parameters is presented. The WACC was decided to 5,2 % for the first regulatory period without any change during the period. Table 3 Rate of return parameters for the first regulatory period WACC parameters Risk free rate nominal 4,0% Market risk premium 4,74% Tax rate 20,0% Asset beta 0,38 Equity beta 0,66 Illiquidity premium 0,50% Debt share 0,50 Debt/equity share 1,00 Expectation on inflation 1,99% Debt premium 1,49% Overview of rate-of-return parameters and the RAB All countries use different approaches when it comes to calculate the RAB. In Denmark the approach is a mix of standard units and book values, norm costs are applied before year 2000 and actual book values thereafter. In Norway only book values are used. In Finland the principle is to calculate an age adjusted replacement value (a proxy for a market value) using standard unit costs. Sweden use replacement value based on standard unit costs. In Table 4 the WACC parameters used in the DSO models in Finland, Norway and Sweden in the present regulatory period are compared. Denmark does not use the WACC approach. The rate-of-return is decided as 30 year bonds on real estate with an addition of 1 %. Both Norway and Denmark apply a nominal WACC to their asset base. This is due to a capital base that is not price adjusted during time. In Finland the value of network is inflation-adjusted annually and therefore real WACC is used. In Sweden the capital base is based on an annuity approach and a real WACC is applied. 33

36 As can be seen from Table 4, the parameters differ between the countries. The capital structure is set differently; where Sweden has a capital structure of 50 % debt and 50 % equity, Finland has defined that 30 % is related to debt, while Norway has 60 % debt in the capital structure. The debt premium also varies a lot between the countries. Table 4 Rate of return parameters for the first regulatory period WACC parameters for DSO Finland Norway Sweden Risk free rate Real : 10 year Finnish government bond yield (average of previous year May) - inflation component Nominal: Annual risk free rate of government bonds with 5 years maturity Nominal interest on government bonds on long run. Estimated to 4 % nominal and 2 % real for a 30 year perspective. Inflation 1,0 % 1,99 Asset beta 0,4 0,35 0,38 Equity beta 0,527 0,875 0,66 Market risk premium 5 % 4 % 4,75 % Premium for lack of liquidity Capital structure (Debt / Equity) 0,5 % 0,5 % 30 / / 40 50/50 Tax rate 26 % 28 % 20 % Debt premium 1,0 % 0,75 % 1, Operating costs Operating cost can be divided in controllable and uncontrollable cost. By the latter means cost which the company has no or little possibility to influence in the short run. This type of cost is often treated as a cost-pass-through variable by the regulation (regulator). Controllable cost is by the regulator treated as possible to influence by the company that is to lower over time for given level of production. The companies can in principle not decide the level of production (in terms of number of customers, demanded power or distributed volume of electricity. The regulation of operating cost differs between the Nordic countries especially if one look at the levels of details. The Nordic countries have a common way to look at the operating cost, but the model for deciding on the level of reasonable cost is different. 34

37 2.5.1 Denmark For each DSO, the necessary costs are defined as the level of operational costs, depreciations and capital costs associated with an efficient operation of the distribution grid in Each year, a DSO s revenue cap is basically determined by the DSO s priceadjusted necessary costs in For each DSO, the costs of grid losses from 2004 were equal to the actual physical losses measured in kwh multiplied by the price of electricity. This cost was fully included in the DSOs initial revenue caps as part of the operational costs. Thus, the costs of grid losses are part of the necessary costs from 2004 which are annually price-adjusted to yield the DSOs revenue caps for each year. If the electricity price increases more than the priceindex which is applied to price-adjust the DSOs revenue caps, then the DSOs receive an increase in their revenue caps. DSO s revenue cap is on the other hand not adjusted for variation in physical grid losses. Thus, a DSO is allowed to keep the entire economic benefit from a reduction of its physical grid losses compared to the physical losses that the DSO had in 2004 and vice versa. This provides the DSOs with an economic incentive to reduce losses Finland In the Finnish regulation model part of the operating costs are considered controllable and they are subject to the efficiency target. Controllable operating costs are defined before start of the regulatory period. See Appendix of the Country report Finland for a list. Network losses are treated as pass-through costs and there is no efficiency target applied on them. DSOs are able to cover some of their investment costs through the Innovation incentive system. The system includes two parts: the expenditure caused by research and development activities (maximum 0,5 % of annual turnover is covered) and the costs for DSOs caused by hourly metering system (5 euros/place of electricity use with max 63A main fuse is covered annually) Norway The controllable operation and maintenance costs mainly comprise of personnel- and material costs related to own production of services and purchase of external services. A grid company can charge a customer for costs related to existing grid when the customer asks for quality or services that are usually not expected to be delivered as a grid service. Revenues following these kinds of services can be kept out of actual revenue, but are also subtracted from the controllable costs that enter the revenue cap. The controllable operating costs are treated as other costs in the revenue cap, where 40 % of all costs are passed through, while 60 % are dependent on the benchmarking results and the cost-norm. The costs related to the network losses are included in the companies costs using yearly reference prices of power. In the benchmarking models, only the costs of network losses in the distribution grid are included. The calculated costs of network losses in regional grid are covered 100 % and excluded from benchmarking, because the loss volume is not only dependent of the network company s decisions, but also influenced by external factors in the grid. 35

38 Other costs treated as non-controllable are tariff costs related to other grids and property tax Sweden Operating cost is the other part of the total cost for the grid services. All operating costs that are defined as controllable will have an efficiency requirement. The starting point is the reports from the DSOs their historical costs taken from audited data from bookkeeping. In theory should the latest year with actual historic data be the starting value for assessing the revenue cap. But to avoid that the year in question can be an extreme year (high or low) the starting value will be set to average of the four years inflated to the price level of Since only controllable costs will be imposed efficiency requirement on one %age each year, it is important to create uniformity as regards the costs should be considered current and capital expenditures. The efficiency target on 1 % is decided upon empirical studies on the productivity development for the last 10 year in the distribution of electricity. 7 The mean growth in productivity was about 2 % each year during these years. The cost that the DSO has no control over is there no efficiency target. The most typical kind of such cost is the charges the DSO has to pay to be connected to TSO (subtransmission). Also the charges the government collet for financing the regulation of the electricity market and the agency for electric security is treated as uncontrollable. The grid losses can from a legal perspective be regarded as controllable. In the short term, however, the DSOs have limited opportunities to influence their costs for grid losses. The physical losses in terms of electricity are not possible to control. The tendering of these losses is possible to control. In the medium and long term the losses are possible to control both by the tendering and by investment in the grid. The grid losses will for the first period be seen as an uncontrollable cost with no efficiency requirement. That is cost for losses are of cost-pass-through type. The control of the tendering process is in the regulation regarded as parallel inspection activity. The starting value for deciding on the operational cost part is the actual cost for the years These four years cost figures are updated to the price level of The mean value for these updated four years historical costs is the input value for the running costs in the first regulatory period Differences and similarities The costs of losses on the grid are treated in different way. In Denmark the cost for losses are treated as a necessary cost and fully included in the revenue caps. A price increase in electricity will increase the revenue cap. The cap is on the other hand not adjusted downwards if the physical losses go down. The DSO gets the whole benefit of less physical losses on the grid and vice verse with increase in losses. Here is an incentive for reducing physical losses. In Finland and Sweden the losses are treated as pass-through-costs and there is no efficiency target on these costs. The losses in Sweden are treated as non-controllable in 7 Energimarknadsinspektionen, Förhandsregleringens krav på effektiviseringar intäktsramen för löpande kostnader, EI R2010:11. 36

39 the calculation of revenue caps. The Electricity Acts states that the electricity for losses is tendered in a competitive way. The auditing that the DSO does buy (tender) the electricity for losses can be done separately as a special activity. In Norway the losses are included in the revenue cap by using yearly reference prices of electricity. If the DSO can buy electricity at a lower price than the reference price they get a benefit. Costs of losses are also included in the benchmarking. The revenue cap is composed by 40 % of the DSO actual costs and 60 % by the cost level of an efficient operating DSO. If the DSO operates efficient due to benchmarking analysis, they will have 100 % of their costs covered. The losses in the regional grid are treated as fully noncontrollable and not part of benchmarking. Operating and maintenance costs for staff and material are treated as controllable in all countries. Efficiency targets are applied on operating costs in all countries. In Sweden there is only a general target on controllable costs, while the other countries use benchmarking methods to set DSO-specific targets. The benchmarking methods used differ between the countries. 37

40 3 Quality regulation Electricity is produced, transported and used simultaneously. Any customer connected to the power system is, to various degrees, influencing and being influenced by the quality of the electricity delivered from the power system. Quality of supply is a collective term which includes dependable delivery, quality of voltage and non-technical elements like customer services and information. In this chapter we limit the quality of supply term to cover only dependable delivery. We describe how interruptions are treated in the economic regulation of the networks in the Nordic countries. Quality of voltage and non-technical elements are not subject to incentive regulation in any of the Nordic countries at the moment Denmark DERA benchmark the DSOs System Average Interruption Frequency Index (SAIFI) and System Average Duration Frequency Index (SAIDI) to assess each company s quality of supply. In Denmark, different types of interruptions are weighted differently in the calculation and benchmarking of SAIFI and SAIDI. Unplanned interruptions are weighted with a 100 % and planned interruptions are weighted with 50 %, Interruptions caused by third party (e.g. a constructing company that damaged a DSOs power cables) are weighted 10 % while Interruptions caused by force majeure are totally excluded from the benchmarking of the DSOs quality of supply. DERA penalises the DSOs with a larger weighted SAIFI than the aggregated DSOs holding 80 % of the distribution grid. Analogously, DERA benchmark the DSOs weighted SAIDI. DSOs are likewise penalised with an up to 1 % reduction in their operational costs if they have a larger weighted SAIDI than the companies with a lower weighted SAIDI holding 80 % of the distribution grid. On aggregate, DSOs with both a relatively high SAIFI and SAIDI can be penalised with an up to 2 % reduction in their operational costs. To avoid that DSOs with a generally high quality in supply neglect some of their customers (e.g. customers living in a scarcely populated part of a DSO s distribution area), DERA punish a DSO if 1 % of the DSO s customers experienced more interruptions during a year than 99,5 % of the aggregate Danish customers experienced during the same year Finland In the annual calculation of the disadvantage caused by outages (DCO) in the electricity supply of network operators, the planned and unplanned outages in long outages and the high-speed and time-delayed auto reconnections in short outages will be taken account. In the calculation, the key figures on the number and duration of outages, describing outages specific for particular network operators and reported to EMV by the network operators, will be used as the outage data. The prices used for various outages are based on the results of a study on the disadvantage of electricity supply outages, which was carried out in

41 The prices of outages obtained as the result of the study have been revised so that, combined with the key figures on outages, gathered annually from the distribution network operators, the cost of electricity supply outages will describe the cost for the customers as accurately as possible. A description of the way the results of the original study have been revised is presented in the end reports of studies commissioned by the EMV. In the third regulatory period, the arithmetic average of the network operator s actual outage costs in adjusted to the value of money for the year in question will be used as the reference value for the network operator s outage costs. This way the reference level represents the normal outage level. When calculating the adjusted profit, DCO will be taken into account in the incentive to improve quality part of the regulatory model. By setting a limit value (so-called floor and ceiling levels) for the difference between the actual DCO in electricity supply and the reference level, EMV takes major deviations into account. A half of the difference between actual disadvantage caused by outages in electricity supply and the reference level of outage costs sets monetary value of the incentive to improve quality. This is the monetary effect of incentive in the calculation of adjusted profit. The value of the incentive to improve quality may correspond with a maximum of 20 % of the reasonable return calculated for capital invested after tax in the year in question. This applies to both floor and ceiling levels Norway In Norway, incentive regulation on continuity of supply is fully integrated in the economic regulation through inclusion of the CENS element (CENS cost of energy not supplied) in the revenue cap. We refer to the country report for Norway for a more detailed description of how this element is handled in the revenue caps and allowed revenues. The CENS element reflects the socio-economic costs related to interruptions, taking into account that interruptions in power supply result in costs for the affected customers. From the regulator s point of view it is important that decisions influencing on the continuity of supply also is based on cost-benefit analyses. I.e. the costs related to measures taken to reduce the extent of interruptions should be lower than the future decrease in customers interruption costs due to the measures taken. That is to achieve an optimal level of continuity of supply for the society as a whole, by giving the grid owners incentives to operate, maintain and develop their grids in a socio-economic optimal way and thereby provide an acceptable level of reliability. The customers costs related to interruptions are detected through nationwide surveys and will vary between different customer groups, when the interruptions occur etc. The costs related to investments to reduce the extent of interruptions will on the other hand depend significantly on the location of the customers connection to the power system, including grid topology, geography, climate etc. 39

42 CENS comprises both notified (planned) and non-notified (unplanned) interruptions. The customers are divided in six groups: industry, commercial, large industry, public, agriculture, residential. Normalized cost data (based on a customer survey conducted in ) are used to establish continuous cost functions for each customer group. In addition to the CENS arrangement, NVE also introduced a direct compensation scheme for very long outages (>12 hours). The standard rates for compensation are specified in the regulation governing financial and technical reporting, income caps for network operations and transmission tariffs. Any end user affected by the outage can claim compensation from the network company. However, the main aim of the scheme is to give the network company incentives to repair any outage as quickly as possible, also in cases where mainly small end users (such as households) are affected Sweden It is not possible to apply a system of interruption statistics per customer for the first regulatory period because such statistics are missing for the moment. Through a brief analysis of the available customer interruption assessments (cost for the customers) as well as by studying the application of quality control in the rest of Europe, the decision by the Swedish regulator will be to use the cost assessment study done by the Swedish Energy (the industry association for DSOs). This study was carried out in Swedish Energy did a simple update of these interrupt costs for customers in 2003 by inflating the values by change in consumer price indices (CPIs). The calculation relating to the outcome of 2012 will therefore be adjusted to the price level in 2012, and so on. In Chapter 5 section 7 of the electricity Act shows that quality deduction in the revenue caps shall be limited to a maximum of an amount corresponding to the return on the capital base. The rule has been added to protect mainly smaller grid companies in case of extreme weather impact. The regulator, EI will therefore limit the annual amount of quality adjustment for the period ( ), by introducing a ceiling and a floor as maximum amounts to return on own funds, but not more than 3 % of annual revenue. The introduction of the ceiling on how high a supplement can be detrimental unjustifiably high supply security while a floor would restrict the negative economic consequences for grid operators. The historical value of interruption produces a reference level for each company. Premium or deduction of any deviation is proportional to the indicators with a ceiling and floor in 3 % of total revenue Overview of quality regulation All countries have included quality of supply in the economic regulation in some form, but the impact of quality of supply in the economic regulation differs between the countries. In Denmark there is a direct link between interruption indexes (SAIFI and SAIDI) and revenues, and the companies are penalised through lowered allowed revenues for poor quality of supply. Finland and Norway both include a calculated cost for customers related to disadvantages caused by outages in the economic regulation. In the Norwegian and Finnish regulation this customer cost is fully integrated in the revenue cap regulation to provide the network operators with incentives to take into account that interruptions also have socio-economic costs when they plan for development and 40

43 maintenance of the grid. Sweden uses indexes SAIFI and SAIDI for setting norm levels from historic levels of outages. These norm levels are compared to the outcome during the regulatory period. The difference between norm and outcome is valued by customer estimated value of lost load. The structure (design) of quality regulation in Finland, Norway and Sweden is closely related. In Finland the realized outage cost is annually compared to the reference level (based on average historical outage cost) with an effect on the adjusted profit. In Sweden there is an adjustment after the regulatory period when reference level is compared to outcome will affect the return on capital. For Norway, a reference level is set through the yearly benchmarking process, comparing one company s interruption costs to the costs of an efficient norm company. The full integration of the quality element further implies that a yearly company specific adjustment is made through the allowed revenues the CENS costs entering the revenue cap is based on interruptions dating two years back, while the actual interruption costs in the revenue cap year is deducted from the allowed revenue for the same year. In Denmark on the other side, the penalty for worse (lacking) quality hits the level of operational cost. Common for all four countries is the special treatment of very long interruptions. Direct schemes for compensating customers who have suffered from very long interruptions are in use. 41

44 4 Differences between regulation of DSO and TSO The same principles for both DSO and TSO are applied as far as possible. The main difference in the regulation between DSO and TSO is due to the fact that the TSO has other responsibilities, as the TSO is appointed the responsibility for the functioning of electricity transmission system on national level. Another difference is that there is only one TSO but many DSOs in each country. This has influenced the design of regulation. 4.1 Denmark The main transmission grid of 400 kv is owned and operated by Energigrid.dk which is the only TSO in Denmark. This grid has a total length of about 6,000 km. Energigrid.dk is regulated according to non-profit regulation. This means that the company can only demand payment to cover necessary costs of efficient operation as well as modest interest on equity capital to secure the real value of its equity. Energigrid.dk tariffs are set on the basis of forecasts of costs for the following year and the volume of electricity transported in the company s grid etc., so revenues and expenses balance in accordance with the non-profit principle. However, the forecasts cannot always be 100% accurate, and the company will therefore either charge too much or too little. This over or under charging is included in the tariffs for the subsequent year. Over coverage will mean lower tariffs in the following year, while under coverage leads to higher tariffs. 4.2 Finland The responsibility on the functioning of the electricity transmission system, a task that distribution grid operators do not have, is the main difference in the legal position of TSO and the DSOs. The regulatory model of TSO is based on same principles as the model for DSOs regulation. The models have several similar building blocks. Regarding the WACCcalculation, there is a difference in the gearing factor. Since the first regulatory period the Energy Market Authority (EMV) has applied to TSO a fixed capital structure of 60/40, i.e. the ratio of interest-bearing debts to equity, which is used in calculating both the beta and the weighted average cost of capital invested in network operations. The selected fixed capital structure corresponds to the average capital structure of TSO when network assets have been adjusted for their net present value. TSO has an efficiency target only on its controllable operational cost (KOPEX). EMV confirms every year of the third regulatory period the starting level of these costs. When measuring the starting level EMV uses KOPEX budgeted by TSO for the year in question or the realized costs of five previous years. Changes in the network volume are taken into account. When measuring the reference level EMV adjusts starting level with the general efficiency target and error margin. The realized KOPEX is subtracted from the reference level annually in the calculation of adjusted profit. 42

45 The incentive to improve quality and the innovation incentive parts of the TSOsregulatory model have also some differences compared to the DSOs regulatory model (see Figure 7, structure of the regulatory model). 4.3 Norway Statnett SF is licensed as the Norwegian transmission system operator (TSO). The role and responsibilities of the system operator follows from the Regulation on system operation, where the embracing goals are to facilitate an efficient electricity market with a satisfactory quality of supply. Statnett responsibilities that does not regard the system operation role, is regulated through the same legislations as other network utilities. The economic regulation is specified in a separate chapter in the regulation governing financial and technical reporting, income caps for network operations and transmission tariffs, where it is stated that the TSO is to be regulated by a joint revenue cap including system operations costs and costs related to own network components. The TSO in Norway is regulated by yearly revenue caps, as any other network company. The main difference is that Statnett is regulated by a joint revenue cap covering also system operations costs. The main principles for determining the revenue cap are the same for the TSO as for any other network company the revenue cap is based on 40 % of Statnett s own costs and 60 % of a cost norm. The company s specific costs entered into the revenue cap for the TSO are calculated based on the same elements as are entered into the DSO s revenue caps. The benchmarking of transmission grid cannot be carried out by the same methods as for DSOs, due to lack of comparability. The benchmarking of the transmission grid is therefore based on the results of the international study on the TSOs, the e3grid study. 8 This study compared 22 TSO in European countries regarding cost efficiency. Statnett also owns part of the regional grid, which is benchmarked by DEA (as for the other network companies), comparing Statnetts data to data for the remaining regional grid companies. 4.4 Sweden EI is under the obligation to propose a decision establishing the revenue cap for Svenska kraftnät (SvK) to the government for the regulatory period in EI will decide on the revenue cap for SvK from A revenue cap should be established in advance of each regulatory period and each regulatory period shall last four calendar years, unless there are special reasons for a different time period. According to 6 of the governmental decree, the regulatory period is one calendar year in regard to revenues from the transmission grid held by SvK. The first difference between DSO and TSO is the regulatory period. The second difference between DSO and TSO is the estimation of the capital cost. The asset base for the TSO is completely calculated from the historical acquisition values for 8 Agrell P and Bogetoft P, e3grid Final Results (2009). 43

46 the physical assets they have in use. The historical values are indexed to account for the inflation to get the cost level of The reason for using acquisition value is the lack of norm values for the TSO assets. 44

47 5 Applied benchmarking The purpose of benchmarking is to get a norm a reference which each company in the studied industry can be compared with. In a standard benchmarking process, according to the literature, a first step is to find the best company which serves as a role-model for the other companies in the industry studied. A second step is to copy the relevant parts of the reference company s management, organization etc. The possibility to use benchmarking in the regulation of infrastructural services depends on the number of independent companies in the industry. In industries that are characterized of a structure with several (or many) local monopolies there is an opportunity to use yardstick competition (or yardstick regulation). Sometimes the term is performance based regulation (PBR). In this chapter we describe the different benchmarking methods applied in the Nordic regulations 5.1 Methods and models During the years after the deregulation, the Nordic countries applied different methods for benchmarking the electricity grid industry. Methods like Data Envelopment Analysis (DEA), Stochastic Frontier Analysis (SFA), semi parametric methods and more have been used for assessing cost efficiency in all countries. In the choice and development of the regulatory models these methods have been applied. By comparing the companies in the industry regarding productivity the regulator gets information on the efficiency structure of the industry. The company with the highest productivity (most cost efficient) can be used as a benchmark for the other companies. The problem with the practical application of different benchmarking techniques as DEA or SFA is the difficulty of finding comparable utilities. To handle the context (environment) the different companies are working under, there is a need for applying some variables to make the comparisons fair. 5.2 Denmark Annually, each DSO has to report its stock of 23 different types of grid components installed in its distribution grid (e.g. kilometres of power cables). DERA has estimated average unit operational costs for each of the 23 different types of grid components. Furthermore, DERA has estimated average unit costs of depreciation for each of the 23 grid components. For each DSO, DERA first multiplies the DSO s stock of each of the 23 different types of grid components with the associated unit operational cost and unit depreciation cost. DERA then sums the 23 products to obtain a measure of the DSO s so called netvolumen. For a given DSO j, the netvolumen measures the cost level that an average DSO would obtain when operating the DSO j s distribution grid. Each DSO also has to reports its total operational costs and total costs of depreciation. DERA applies the reported costs to calculate a cost-index for each DSO: 45

48 According to DERA (2009), operational costs are higher in densely populated areas than in scarcely populated areas. DERA therefore adjust the cost-index for variation in population density among the DSOs. The DSOs with a relatively low adjusted cost-index are relatively cost-efficient and vice versa. DERA applies an average of the top 10 % most cost-efficiency DSOs to benchmark the cost-efficiency among the remaining 90 % of the Danish DSOs. Based on this benchmarking, DERA sets an annual efficiency requirement for each of the DSOs with a relatively low cost-efficiency compared to the top ten most cost-efficient DSOs. The efficiency requirements imply that each of the DSOs with a relatively low cost-efficiency has to become at least as cost-efficient as the average of the top ten most cost-efficient DSOs within a five year period. The benchmarking that DERA e.g. performed in 2009 was based on data for For each DSO with a relatively low cost-efficiency, DERA has required that the DSO reduces the difference in cost-efficiency to the average of the top 10 most cost-efficient DSOs in 2008 with 1/5 during DERA is going to perform a new benchmarking during 2010 which is going to be based on data for Once again, DERA will require that the DSOs with a relatively low cost-efficiency each reduce the difference to the average of the top ten most cost-efficient DSOs in 2009 with 1/5 during Thus, the efficiency requirements are calculated annually. 5.3 Finland The incentive to improve efficiency contains an efficiency target that includes a companyspecific efficiency figure (estimated for each DSO) And a general efficiency target (same for each DSO). The methods used for earlier regulatory periods have been Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA). For the third regulatory period has the Energy Market Authority applied a new method for the benchmarking of the DSO called StoNED (Stochastic Non-smooth Envelopment of Data). EMV has applied the general efficiency target since the first regulatory period. The yearly target of 2,06 %, based on general productivity and technological progress was initially estimated in The figure was estimated with updated data in The most recent study suggested a figure very close to the target for the first regulatory period, so EMV found no reason to change the size of general efficiency target. The estimation of company specific efficiency target uses data on each DSOs input, outputs and contextual variables from years Average for the year of total costs of each DSO is indexed into year 2010 value of money. Total cost is a combination of the actual operational cost and half of the actual disadvantage caused by electricity supply outages to the customers of the DSO i in the year t. The indexation is done by consumer price index. The model contains three outputs: amount of energy distributed (GWh), length of the company s network (km) and number of customers connected to the grid. The amount of energy distribution captures the direct (variable) output of the distribution activity. The amount of energy distribution is weighted with the average national 46

49 distribution prices of years (0,4 kv, 1-70 kv and 110 kv). Using fixed weights ensures that comparison between different years will not be affected by inflation or changes in the electricity. The length of network and the number of customers represent the potential output or the capacity. These variables capture the fixed cost of maintaining a sufficient capacity to serve their designated network area irrespective of the actual consumption of power. While the output variables two and three can draw a distinction between urban versus rural networks, the networks located in large cities have very similar output structure as those located in suburbs or small towns. To better capture the differences in the output structures and operating environments of the urban versus suburban networks, a contextual variable has been placed into the model. The proportion of medium voltage (1-70 kv) underground cables is used as the contextual variable. The proportion is calculated based on the total network length (average of years ). In order to get the company-specific efficiency target, the estimated efficiency of each DSO and the general efficiency target are combined. The transition period of eight years is taken into account in the calculation of the company specific efficiency target. The transition period has been set because differences in efficiency between DSOs appear to be wide and the cost adjustment into efficient level can be softened for the DSOs with high potential for efficiency improvements. 5.4 Norway NVE benchmark the network companies with the DEA-method. Two separate models have been constructed: one model for the distribution grids and one model for the regional and central grids (but without Statnett). The input variable in both models shall reflect the total use of resources of a company, which are given by their total costs. Common for all grid levels, is that annual depreciations and return on capital related to grid financed through contribution are included in the costs. NVE has defined a set of output variables for the two models that shall describe the most relevant cost-drivers. Total costs are minimized in the analysis, given these outputs. NVE has arrived at the variables by extensive testing of different relevant variables. There are 8 cost-drivers in the model for distribution grid and 5 in the model for regional- and central grid. Some of the outputs describe the companies costs related to their structural conditions (lines and cables, transformers etc.), some are a measure of customer demand (e.g. energy delivered, no of customers) and some are constraints as geographical conditions (e.g. forest, snow). They are all cost-drivers to grid companies. The output variables for regional and central grid: Weighted value of overhead lines Weighted value of underground cables Weighted value of submarine cables Weighted value of components in stations (switches, transformers and compensators) Forest 47

50 The output variables for distribution grid: Energy Demand (MWh) No. Customers ex holiday cottages No. Holiday Cottages Km grid above 1kV No. Transformers Forest Snow Coastal Exposure (wind and coast) From 2010, NVE introduced a regression stage in the model for distribution grid. The DEA results from the first analysis are adjusted through a regression analysis. The coefficients in the regression are a result of a panel data model, where the DEA results define the dependent variable and the variables listed below defines the independent variables. One of the reasons for introducing this second stage is that a DEA model does not work well when too many variables are included. Three variables are included in this second stage: Interface to regional grid (components in station) MW Input from small hydro power stations No. islands 1 km or more from land or another supplied island NVE has developed a weight system that describes the relative relation between the costs connected to each component that comes into the variables overhead lines, underground cables, submarine cables and station components included in the model for regional and central grid and in the variable interface to regional grid in the regression stage in the model for distribution grid. The system refers to both capital - and O&M costs and contains 185 components of overhead lines, 44 components of underground cables, 34 components of submarine cables, eight types of switches, five types of transformers, transformer capacity and six types of compensators. Several geographical variables are included in the model. Differences in geographical and climate conditions between companies may give different cost levels. The conditions that are included in the models today are forest, snow and coastal exposure. Two separate analyses are carried out in each of the models. In the first analysis, the companies are benchmarked using data from two years back. The results from this analysis rank the companies efficiency from 0 to 1 using the most recent data available. The efficient companies get a DEA result equal to 1. To rank the efficient units from this analysis, we make use of historical data from the last 5 years available in a second DEA analysis. The efficient companies are here benchmarked to see if they have become more efficient, e.g. spends less on the same deliverables, than they did on average the last 48

51 5 years. The companies are rewarded with a DEA result larger than 1 if the analysis show improvement, if not, they get a DEA result equal to Sweden For the first regulatory period with ex ante regulation there is only a general X-factor applied for the revenue caps. In other regulatory regimes there is a system of both a general and company-specific X-factor. For the first period the regulator has decided to set X to 1 %age per year from 2010 to That is, the revenue cap will each year be reduced by this %age of the operational cost which the companies have possibility to reduce by being more efficient. An analysis of the productivity development for the years in the DSO sector in Sweden was done with application of both regression analysis, SFA and DEA. The mean development of productivity was estimated to 2% per year for this period. Other studies of the productivity development were also used as information. The model used for the estimation of productivity development consisted of controllable operative cost as input and three outputs (number of customers, length of lines and cables and installed capacity of transformers. There was no need to incorporate contextual variables because the focus was to estimate the development of productivity and not to make comparisons between the companies. Earlier models for bench marking used to incorporate the volume of electricity distributed, which is common in other models. The reason for not including distributed electricity is that the grid losses in not included in the operating cost. 5.6 Different methods and use Benchmarking in the regulation is used for setting efficiency targets for the DSO and TSO. The Nordic regulators have applied different approaches for that purpose. In Norway the regulator has used Data Envelopment Analysis (DEA) for setting efficiency targets since 1998 in the calculation of revenue caps. In 2010 they also introduced a second stage in the model for DSOs, were the DEA-scores are adjusted in a regression to take into account differences in structural conditions between the companies. EMV has continuously developed the efficiency benchmarking in the regulatory model of DSOs. For the 2nd regulatory period SFA-method was introduced to supplement the DEA-method. For the 3rd regulatory period these methods are replaced with a less restricted, semi-parametric StoNED-method (Stochastic Non-smooth Envelopment of Data). In fact, both DEA and SFA could be obtained as constrained special cases of the more general StoNED-method. In Denmark the regulator has chosen a benchmarking method built on cost norms for the regulation. For a given DSO, this calculated cost measures the cost level that an average DSO would obtain when operating the DSO s distribution grid. The benchmarking level is then applied as an average of the top 10 % most cost efficiency DSOs to benchmark the cost efficiency among the remaining 90 %. Based on this benchmarking is the annual efficiency requirement set The Swedish regulator has decided to use a general target for all grid operators. This is different compared to Finland that use both a general and a company specific target and to Denmark and Norway which only use a company specific target. 49

52 There is also a difference in the choice of model for measuring the efficiency and productivity development between the countries. The Norwegian DSO model has comparably many variables with one input (total cost) and eight outputs (cost drivers) in the DEA-model and in addition three variables in the regression step. The model used in Finland has also total cost as input, but only three outputs plus one contextual variable, in total five variables. 50

53 6 Evaluation of regulation and issues in future regulation Because the principles and methods of the economic regulation apply to regulatory periods in the Nordic countries, the possibility to evaluate will differ between the countries. There is always a time-lag between a change in a regulatory model and the impacts this change can have on a single company and for the industry as a whole. Even when the founding principles are constant for several regulatory periods, there are usually some minor changes which in some way change the model. These minor changes are often a result of evaluations. Evaluations of a regulation can also be done at different levels. From studies of the development of productivity in the industry, to evaluations of administrative processes in the regulatory agency and the industry. In this chapter we also address the future regulation. What will be the future possible changes in the current regulation? The plans for future regulation have different status in terms of when they will be implemented or if they will be implemented at all. 6.1 Finland On the request of the Energy Market Authority, Lappeenranta University of Technology conducted an evaluation on the Finnish electricity distribution model. The objective of the project was to evaluate functionality and steering effects of the regulatory model. Based on these the project group should assess development needs of the model. The report was part of the Roadmap 2020 project. EMV started Roadmap 2020 project to find out megatrends of near future in the electricity distribution field and develop the regulatory model The project group studied regulatory model by evaluating data of grid operators. The report included also a survey and network operators were interviewed about their opinion on the different parts of the regulatory model. The incentivizing effect of the regulatory model was analysed. The report was conducted in 2010 and it focused on the second regulatory period. Overall the regulatory model was found to function well. A few development ideas were suggested and EMV has taken some of these suggestions into account when preparing the model for the third regulatory period EMV started a Roadmap 2020 project. The objective of this project was to find out the megatrends in electricity distribution in the near future, develop the regulatory model and define a strategy for regulation for 2020 and onwards. There were four megatrends in network industry identified in the project: 1. Significant need for replacement investments Network lifetimes are coming to end The arrival of smart meters and smart networks DSOs will concentrate on their core competence and strategy 2. Society s increased reliance on electricity 51

54 Interruptions will not be accepted Exceptional storms will happen more often 3. Emission free electricity production increases Increased energy production using wind farms and bio-fuels Two-way distribution and micro production units More nuclear power in networks 4. Integrated regulation Detailed steering and regulation at EU level (Acer) It is possible that above issues have to be concerned in the future development of the regulatory model. 6.2 Norway In 2010, the Ministry of Petroleum and Energy (MPE) had carried out an external, independent evaluation of the economic regulation of the Norwegian grid companies. This evaluation was performed by Nils-Henrik M. von der Fehr, professor at Department of Economics, University of Oslo. The main conclusions from this evaluation indicate that the economic regulation is based on approved principles, is leading in an international perspective, and is developed over time with adjustments based on experience and new challenges. The evaluation does support the economic regulation in high degree, but point at some challenges. Further, NVE has carried out several evaluations of different aspects of the economic regulation. Some have been performed by use of external consultants. Others have been performed internally, based on own expertise in combination with discussions with the electricity network industry. A major evaluation was performed during 2010, in parallel and responding to the evaluation carried out by the MPE. Since the current revenue cap regulation came into force in 2007, NVE has had dialogue with the network sector regarding the companies incentives to make necessary investments. In NVEs view, the main incentives to invest should be (and are) provided through direct regulations, and not economic regulation, as described in chapter 1. However, NVE acknowledge that the grid sector find the cost norm model challenging. This challenge is mainly due to the implications of the model for cash flow from investments, as the existing model implies that a substantial part of the cash flow will be recovered late in time, as a consequence of the fact that capital costs are entered into the DEA analyses using book values and linear depreciations. This challenge was also pointed out by Professor von der Fehr. NVE is currently investigating the possibilities for accelerating the cash flow from investments, through a modification of the cost norm model. The actions under consideration do not require any major changes in the regulation. Some adjustments will therefore enter into force already from The quality of the data that is entered into the DEA analyses has also been subject for discussion over the past years. Challenges are connected to two aspects; both consequences of variation in the data for a single company, and consequences of 52

55 possibilities of faults in the specification of the DEA model. Both Professor von der Fehr and the utilities have pointed out challenges in this respect. NVE is also considering adjustments in the cost norm model to address these issues. Some adjustments will enter into force already from 2012, others from In the period after the introduction of the Energy Act in 1991, the main focus was market orientation and efficient utilization of the existing network. These elements are still important, but in addition there is an increasing attention towards the need of new- and reinvestments in the transmission system. This is regarding both an increase in the production of renewable energy and the need for re investments in the existing electricity network. The principles laid down in the Energy Act are not amended to facilitate such a priority, but some adjustments in recent years have been done. Among these; requirements for the network companies to connect all new production to the grid, if they are socioeconomical beneficial projects, replacement of all meters with smart meters within the end of 2016, increased economic incentives to prevent interruptions by introduction of direct payment to costumers that experience disconnections lasting more than 12 hours and by implementing short interruptions (less than 3 min.) in the economic regulation. A continuous evaluation of the economic regulation is undertaken to make sure that adjustments according to new requirements are done when it is found necessary. There is an increased focus on smart grids also in Norway, but the focus might not be as large as in other European countries, in the sense that the existing grid structure has already been built to take care of variable generation as hydropower. NVE will aim at increasing its activity regarding monitoring the network companies compliance with the new requirements and direct regulations. 6.3 Sweden The new regulation working from 2012 there can of course be no evaluation. For the years before there have been some evaluation. The Swedish regulator commissioned a consultancy study on the net performance assessment model (NPAM). 9 An empirical evaluation of NPAM is also done as a research project. 10 Because the application of this model was appealed to administrative court when the model was implemented, there have been several expert reports delivered to the court from the appealing companies to the court. Another study of the DSO industry in Sweden was done in a dissertation. Different aspects on the distribution industry were studied. 11 Because of the decisions of revenue caps in 2011 the focus is entirely on the first regulatory period. The work with the actual regulatory model has naturally incorporated issues for how to regulate in future periods. With capital cost calculated with real 9 Jamasb T & Pollit M, Reference Models and Incentive Regulation of Electricity Distribution Networks: An Evaluation of Sweden`s Network Performance Assessment Model (NPAM), September 2007 CW0747&EPRG Jamasb T & Söderberg M, The Effects of Average Norm Model Regulation: The Case of Electricity Distribution in Sweden, Review of Industrial Organization, Söderberg M, Four essays on efficiency in Swedish electricity distribution, diss Gothenburg University. 53

56 annuities on replacement values there is a discussion on develop norm cost data for operating cost for each type of physical grid assets. 54

57 7 References Agrell P and P Bogetoft, e3grid Final Results (2009). Damsgaard N and Green R, Den nya elmarknaden framgång eller misslyckade?, SNS förslag 2005 Energimarknadsinspektionen 2009, Förhandsreglering av elnätsavgifter principiella val i viktiga frågor. Energimarknadsinspektionen, Förhandsregleringens krav på effektiviseringar intäktsramen för löpande kostnader, EI R2010:11. Jamasb T & Pollit M, Reference Models and Incentive Regulation of Electricity Distribution Networks: An Evaluation of Sweden`s Network Performance Assessment Model (NPAM), September 2007 CW0747&EPRG0718. Jamasb T & Söderberg M, The Effects of Average Norm Model Regulation: The Case of Electricity Distribution in Sweden, Review of Industrial Organization, Shleifer A, 1985, A theory of yardstick competition, Rand Journal of Economics. SOU 2007:99, Förhandsprövning av nättariffer m m, Söderberg M, Four essays on efficiency in Swedish electricity distribution, diss Gothenburg University. Article 37 1a, Directive 2009/72/EC of the European Parliament and of the Council of 13 July 2009 concerning common rules for the internal market in electricity and repealing Directive 2003/54/EC. The Agreement on the European Economic Area (EEA). Se EFTA homepage: 55

58 8 Country report: Denmark 8.1 Number of DSOs in Denmark In Denmark, we have one main transmission system operator called Energinet.dk, 8 regional transmission system operators and 84 distribution system operators (henceforth DSO) in Denmark. Energinet.dk is responsible for the national 400 kv transmission network. In some Danish regions, Energinet.dk is furthermore responsible for the 150 kv transmission network. The regional transmission system operators each operate the 150 kv transmission network in one of the remaining 8 regions in Denmark. There is a significant variation in size, ownership-structure etc. among the Danish DSOs. The largest DSO DONG Energy Distribution has more than a million customers which is more than 20 % of the Danish population while the smallest DSOs only have a couple of thousand customers of the Danish DSOs are relatively large compared to the remaining Danish DSOs. These DSOs together serve a very large part of Denmark. Many of the remaining DSOs are located as small islands inside the larger DSOs distribution network. A relatively large DSO called SEAS-NVE Net e.g. has five very small DSOs located as small islands inside its distribution network. During the last couple of years, there have been relatively many mergers among the Danish DSOs. Thus, the number of Danish DSOs is rapidly declining General description of the Danish revenue cap regulation In Denmark, we apply revenue cap regulation and a maximum rate of return on network assets to control the DSOs revenues in Denmark. If a DSO does not violate its revenue cap and the maximum rate of return on network assets, the DSO is free to set its distribution tariffs as long as the DSO does not discriminate between its different types of customers (e.g. between private consumers and industrial companies). Some DSOs choose not to fully exploit their revenue caps. Among other things, this has created a large variation in distribution tariffs among the DSOs. A DSO s capital costs, depreciations and operational costs are aggregately termed necessary costs in Denmark as illustrated in Figure 1 below. 56

59 Figure 1 Determination of the revenue cap DKK Adjustment for inflation Necessary cost Capital costs Depreciations Operational costs Year For each DSO, the necessary costs are defined as the level of operational costs, depreciations and capital costs associated with an efficient operation of the distribution network in Each year, a DSO s revenue cap is basically determined by the DSO s price-adjusted necessary costs in 2004 as illustrated in Figure 1 above. In a given year, a DSO s revenues cannot exceed the DSO s price adjusted necessary costs from However, the Danish Energy Regulatory Authority (DERA) can increase a DSO s revenue cap if the company performs a necessary investment or in case of a significant increase in the costs of network losses due to rising electricity prices as explained in section 5 below. A necessary investment is defined as an investment that either improves a DSO s quality of supply or expands the distribution network to include new customers or new generation facilities. A DSO s revenue cap is permanently increased with the increased costs of network losses while the DSO s revenue cap is only temporarily increased with some allowances due to a necessary investment as shown in Figure 2 below. 57

60 Figure 2 Connection between necessary investments, network losses and the revenue cap DKK Adjustment due to rising costs of network losses Allowances due to necessary investments Adjustment for inflation Capital costs Depreciations Operational costs Year Figure 2 illustrates a constructed example of a DSO that has increasing costs of network losses and performs a necessary investment in For a given year, the faded striped area illustrates the revenue cap from the preceding year which is price-adjusted with the blue area placed on top of the faded striped column. In 2006, the DSO had increasing costs of network losses due to rising electricity prices. For 2006, the DSOs revenue cap is therefore increased with the black area as illustrated in Figure 2 above. If the DSO s costs of the network losses in the subsequent years does not change significantly as explained in Section 5, the associated increase in the DSO s revenue cap due to the increased costs of network losses in 2006 will be price-adjusted like the necessary costs from 2004 as illustrated in Figure 2 above. A DSO has to report the construction costs of necessary investment to DERA. Currently, all necessary investments are depreciated straight-line over a 40-year period. DERA has estimated unit operational costs for a 40-year period for each of the different types of necessary investments. The sum of the constructions costs and the by DERA estimated operational costs of a necessary investment is called the total necessary network expense in Denmark. The total necessary network expense is straight-line depreciated over a 40-year period. The DSO s revenue cap is annually increased with the annual depreciation plus a return on the necessary investment. The allowed rate of return is equal to the 30-year interest rate on mortgage bonds plus 1 %. According to DERA (2010), the interest rate on 30-year building bonds plus 1 % was equal to 6,1 % in A necessary investment completed in 2006 is then allowed an annual return of 6,1 % on the necessary network expense minus depreciations during the entire 40-year period. The annual increase in a DSO s revenue cap is illustrated by the green areas in Figure 2. These areas decrease over the 40-year period due to the annual depreciation of the necessary network expense. In 2045, the necessary network expense has been depreciated to zero. Accordingly, the DSO s revenue cap is then equal to the 58

61 price-adjusted necessary costs from 2004 and the price-adjusted increase in the DSO s revenue cap in 2006 due to increasing costs of network losses. However, DERA can also decrease a DSO s revenue cap in case of low cost-efficiency and/or low quality of supply. DERA annually performs a benchmarking of the DSOs cost-efficiency and quality of supply. If a DSO has a low quality of supply and/or a low cost-efficiency in e.g. 2010, DERA then penalise the DSO with a reduction in the DSO s revenue cap in The benchmarking technique applied to determine the cost-efficiency of the DSOs is described in Section 2 below. 59

62 8.2 Benchmarking of the DSOs cost efficiency Annually, each DSO has to report its stock of 23 different types of net components installed in its distribution network (e.g. kilometres of power cables). DERA has estimated average unit operational costs for each of the 23 different types of net components. Furthermore, DERA has estimated average unit costs of depreciation for each of the 23 net components. For each DSO, DERA first multiplies the DSO s stock of each of the 23 different types of net components with the associated unit operational cost and unit depreciation cost. DERA then sums the 23 products to obtain a measure of the DSO s so called netvolumen: where: 23 net _ component i 1 i, j ( Uope Udep ) netvolumen j = = i + netvolumen j equals the sum of the 23 products for DSO j. net_component i,j equals DSO j s stock of net component i. Uope i = Unit average operational cost for net component i. Udep i = Unit average depreciation cost for net component i. i For a given DSO j, the netvolumen measures the cost level that an average DSO would obtain when operating the DSO j s distribution network (stock of the 23 net components). Each DSO also has to reports its total operational costs and total costs of depreciation. DERA applies the reported costs to calculate a cost-index for each DSO: Cost - index j = ( Operational cost + Depreciation of capital ) j Netvolumen j j According to DERA (2009), operational costs are higher in densely populated areas than in scarcely populated areas. DERA therefore adjust the cost-index for variation in population density among the DSOs. The DSOs with a relatively low adjusted cost-index are relatively cost-efficient and vice versa. DERA applies an average of the top 10 % most cost-efficiency DSOs to benchmark the cost-efficiency among the remaining 90 % of the Danish DSOs. Based on this benchmarking, DERA sets an annual efficiency requirement for each of the DSOs with a relatively low cost-efficiency compared to the top ten most cost-efficient DSOs. The efficiency requirements imply that each of the DSOs with a relatively low cost-efficiency has to become at least as cost-efficient as the average of the top ten most cost-efficient DSOs within a five year period. The benchmarking that DERA e.g. performed in 2009 was based on data for For each DSO with a relatively low cost-efficiency, DERA has required that the DSO reduces the difference in cost-efficiency to the average of the top 10 most cost-efficient DSOs in 60

63 2008 with 1/5 during DERA is going to perform a new benchmarking during 2010 which is going to be based on data for Once again, DERA will require that the DSOs with a relatively low cost-efficiency each reduce the difference to the average of the top ten most cost-efficient DSOs in 2009 with 1/5 during Thus, the efficiency requirements are calculated annually. There is only a weak connection between a DSO s revenue cap and the efficiency requirement from the benchmarking. During 2010, DERA is going to set efficiency requirements for 2011 based on the DSOs operational costs from However, the DSOs revenue caps for 2011 are mainly determined by the DSOs price-adjusted necessary costs from 2004 as stated in Figure 1 above and Figure 3 below. Figure 3 Benchmarking of the DSOs operational costs DKK Revenue cap for 2009 = Price-adjusted costs from 2004 Costs of capital Depreciations Costs applied in the benchmarking for 2010 Operational costs År Year If a DSO had temporarily higher necessary costs in 2004 than in 2003 and from as illustrated in Figure 3, then the DSO currently has a relatively high revenue cap compared to its actual necessary costs. However, since the DSOs necessary costs are significantly lower in e.g than in 2004, the DSOs could appear relatively costefficient according to DERA s benchmarking. On the other hand, if a DSO had temporarily lower necessary costs in 2004 than in 2003 and from , then the DSO currently has a relatively low revenue cap compared to its actual necessary costs and could at the same time appear relatively cost-inefficient in DERA s benchmarking. The revenue caps limit the DSOs distribution tariffs. Due to the weak connection between the DSOs revenue caps and the efficiency requirements, some DSOs with relatively high revenue caps and thus relatively high tariffs, could receive relatively low efficiency requirements. On the other hand, some DSOs with a relatively low revenue cap and thus low tariffs could also receive relatively high efficiency requirements. 61

64 8.2.1 Measuring quality of supply Annually, DERA benchmark the regional DSOs quality of supply. More specifically, DERA benchmark the DSOs System Average Interruption Frequency Index (SAIFI) and System Average Duration Frequency Index (SAIDI) to assess each company s quality of supply. In Denmark, different types of interruptions are weighted differently in the calculation and benchmarking of SAIFI and SAIDI. Unplanned interruptions are weighted with a 100 %, Planned interruptions are weighted with 50 %, Interruptions caused by third party (e.g. a constructing company that damaged a DSOs power cables) are weighted 10 % while Interruptions caused by force majeure are totally excluded from the benchmarking of the DSOs quality of supply Benchmarking quality of supply among Danish DSOs Aggregate level DERA penalises the DSOs with a larger weighted SAIFI than the aggregated DSOs holding 80 % of the distribution network. The example stated in Table 1 below illustrates how DERA identify regional DSOs with a relatively poor quality of supply. The example includes five DSOs: A, B, C, D and E. Table 1 Calculation of the threshold value for poor quality of supply Company A B C D E Total Interruptions (Weighted SAIFI) Km of transmission network cables Share of the aggregate transmission network 8.9% 31.1% 13.3% 26.7% 20.0% 100.0% Accumulated share the Danish regional transmission network 8.9% 40.0% 53.3% 80.0% 100.0% Source: DERA (2009) Company A has the lowest weighted SAIFI while Company B has the second lowest and so forth. Together, Company A, Company B, Company C and Company D have precisely 80 % of the aggregate transmission network as illustrated in Figure 4 below. Figure 4 Identification of regional DSOs with poor quality of supply Interruption per customer 0.11 Company E 0.1 Company D Company C Company B Company A % 20% 40% 60% 80% 100% Share of aggregate regional transmission network % Source: DERA (2009) 62

65 Company D has a weighted SAIFI of 0,09. Thus, companies which have a weighted SAIFI higher than 0,09 are penalised with an up to 1 % reduction in their allowed operational costs. In this example, Company E is penalised. Analogously, DERA benchmark the DSOs weighted SAIDI. DSOs are likewise penalised with an up to 1 % reduction in their operational costs if they have a larger weighted SAIDI than the companies with a lower weighted SAIDI holding 80 % of the distribution network. On aggregate, DSOs with both a relatively high SAIFI and SAIDI can be penalised with an up to 2 % reduction in their operational costs Benchmarking of worst served customers To avoid that DSOs with a generally high quality in supply neglect some of their customers (e.g. customers living in a scarcely populated part of a DSO s distribution area), DERA punish a DSO if 1 % of the DSO s customers experienced more interruptions during a year than 99,5 % of the aggregate Danish customers experienced during the same year. E.g. if 99,5 % of the Danish customers experienced from 0 to 1 interruption during a year, then a DSO will be punished if the 1 % of the DSO's customers with most interruptions experienced 2 interruptions (or more) during the same year. In 2010, DERA for the first time applied the previously described penalising mechanism to improve quality of supply for the worst served customers. Thus, we do not jet know whether or not this mechanism is working as intended. However, DERA has so far only penalised 3 distribution companies out of 85 distribution companies for having an unacceptably low quality for the worst served customers. DERA compares the weighted interruption frequency for each company's 1 % worst served customers with the weighted interruption frequency for 99,5 % of all Danish customers. Thus, if 99,5 % of the Danish customers experienced between 0 and 1 weighted interruption per year, a company is then penalised if the 1 % of the company's customers who were the worst served customers experienced more than 1 weighted interruption per year. Unplanned interruptions are weighted 100 %, Planned interruptions are weighted 50 %, Interruptions caused by third party are weighted 10 % while interruptions due to force majeure are not included in the analysis. 8.3 Regulation of distribution network tariffs In Denmark, a distribution company is free to set its own tariffs as long as the company does not violate its revenue cap and allowed rate of return and, furthermore, does not discriminate among different types of end customers e.g. industrial companies and private consumers. However, most distribution companies apply a standard method developed by the Danish Energy Association to set tariffs. This standard method has been approved by DERA. In Denmark, end customers are typically divided into three groups: Large industrial companies, small companies and private consumers. The large industrial companies are supplied on a higher voltage level than the smaller companies while the private consum- 63

66 ers are supplied on the lowest voltage level. If a distribution company applies the standard method to set tariffs, then the costs of supplying the large industrial companies on the high voltage grid are shared equally among the three different types of consumers. The costs of supplying electricity to the smaller companies on the medium voltage grid are shared equally among the small companies and the private consumers while the costs of supplying the private consumers on the low voltage grid are solely paid by the private consumers. Thus, the distribution tariff that the private consumers has to pay equals 1/3 of the costs of running the high voltage grid, ½ of the costs of running the medium voltage grid and finally the total costs of running the low voltage grid. This method is reasonably transparent. However, some costs are hard to divide among different voltage levels like e.g. administrative costs. Furthermore, some distribution companies are owned by a company that also owns a retail sales company. The retail sales companies often buy resources from the distribution company. According to Danish law, the transactions between these companies must be market based. However, it is not always clear what a market based transaction is Variation in distribution tariffs Some of the DSOs are owned by the municipality in which they are located while others are privately owned. The largest DSO DONG Energy is mainly owned by the state. The privately owned DSOs are either owned by the DSOs customers who generally want to keep tariffs as low as possible while other DSOs are owned by large companies that set tariffs in order to maximise profits. Accordingly, distribution tariffs vary significantly among the DSOs. Furthermore, variation in geographical conditions among the Danish DSOs might also explain some of the variation in the DSOs network tariffs in Denmark Development in distribution tariffs For Denmark, this is a very hard question to answer. For some of the Danish distribution companies, the network tariffs have increased more than 100 % from 2005 to 2009 while other DSOs tariffs have increased with less than the inflation rate. 8.4 The regulatory asset base The initial regulatory asset base was determined in For a large number of net components, DERA estimated unit investment costs and life-expectancy for each of the net components. The initial regulatory asset base for a DSO was then calculated by multiplying the DSO s stock of each net component with the associated estimated unit investment costs adjusted for depreciation. Each net component is depreciated straight line. Consider a DSO that in year 2000 only had 10 units of a certain net component. Each unit is five years old. The net component has a unit investment cost of a 100 Euros and its life-expectancy is 40 years. The initial regulatory asset base for the DSO is then equal to: 10 (100 Euros 5 years (100 Euros/40 years)) = 875 Euros Investments that have been made after the year 2000 are added to the initial regulatory assets base with the actual investment cost. Thus, the unit investment costs were only applied to value net components installed before year The estimated lifeexpectancies for the different types of net components are, however, applied to straight 64

67 line depreciate both the DSOs network assets installed before year 2000 and from 2000 and onwards. The regulatory asset base for a DSO in 2001 is thus equal to the initial regulatory asset base from 2000 minus depreciations plus the actual costs of new investments made in The regulatory assets base for a DSO in 2002 is then equal to regulatory asset base from 2001 minus depreciations plus the actual costs of new investments made in 2002 and so forth Allowed rate of return The DSOs revenues are furthermore regulated with a maximum return on network assets which each year is determined by the interest rate on the 30-year building bond plus 1 %. Unlike in many other countries, we thus do not apply the Weighted Average Cost of Capital (WACC) to determine the allowed rate of return on capital. A DSO s annual result from operating its distribution network divided by the DSO s regulatory asset base may not exceed the interest rate on the 30-year mortgage bond plus 1 %. According to DERA (2010), the interest rate on 30-year mortgage bonds plus 1 % was equal to 7,5 % in Accordingly, a DSO s annual result from operating its distribution network divided by the DSO s regulatory asset base had to be less than 7,5 % in As previously explained, the regulatory asset base is determined annually. Since it is only the price-adjusted costs of capital and depreciations from 2004 that are included in a DSO s revenue cap, there is only a weak connection between a DSO s revenue cap on the one hand and the maximum return on network assets and the DSO s regulatory asset base on the other hand Violation of the allowed rate of return If a DSO violates the allowed rate of return during a certain year, then the DSO has to pay the violation back to its end customers during the following year. Consider the following example. A DSO violates the allowed rate of return with DKK during Annually, DERA audits and approves the DSOs annual account for the previous year. Consequently, DERA will discover the DSO s violation of the allowed rate of return during According to Danish legislation, the DSO must pay the violation back to its end customers during the two subsequent years after DERA has discovered the violation. Accordingly, the distribution company must pay the violation back to its end customers during 2012 and Due to Danish law, the DSO must pay back one third of the violation during the first year (2012) and two thirds of the violation during the second year (2013) Allowed profits A DSO s allowed profits are determined as stated below: Allowed profits = Allowed rate of return Regulatory asset base 8.5 Network losses For each DSO, the costs of network losses from 2004 were equal to the actual physical losses measured in kwh multiplied by the price of electricity. This cost was fully included in the DSOs initial revenue caps as part of the operational costs. Thus, the costs of network losses are part of the necessary costs from 2004 which are annually priceadjusted to yield the DSOs revenue caps for each year. If the electricity price increases more than the price-index which is applied to price-adjust the DSOs revenue caps, then 65

68 the DSOs receive an increase in their revenue caps as illustrated in Figure 2 above. A DSO s revenue cap is on the other hand not adjusted for variation in physical network losses. Thus, a DSO is allowed to keep the entire economic benefit from a reduction of its physical network losses compared to the physical losses that the DSO had in On the other hand, a DSO has to bear the entire additional costs associated with an increase in the DSO s physical losses compared to This provides the DSOs with an economic incentive to reduce losses. 8.6 Regulation of the TSO The main transmission grid of 400 kv is owned and operated by Energinet.dk Denmark s only TSO. This grid has a total length of about 6,000 km. Energinet.dk is regulated according to non-profit regulation. This means that the company can only demand payment to cover necessary costs of efficient operation as well as modest interest on equity capital to secure the real value of its equity. Energinet.dk tariffs are set on the basis of forecasts of costs for the following year and the volume of electricity transported in the company s grid etc., so revenues and expenses balance in accordance with the non-profit principle. However, the forecasts cannot always be 100 % accurate, and the company will therefore either charge too much (over coverage) or too little (under coverage). This over or under coverage is included in the tariffs for the subsequent year. Over coverage will mean lower tariffs in the following year, while under coverage leads to higher tariffs. 66

69 9 Country report Finland 9.1 The objectives of regulation According to EU legislation (Directive 2003/54/EC), national regulatory authorities are to ensure that transmission and distribution tariffs are non-discriminatory and cost-reflective. In addition, the directive further states that the distribution system operator shall maintain a secure, reliable and efficient electricity distribution system in its area with due regard for the environment. And finally, distribution tariffs should be sufficient to allow the necessary investments in the networks to be carried out in a manner allowing these investments to ensure the viability of the networks. According to Section 38 a 1 of the Electricity Market Act, the Energy Market Authority shall confirm for every network operator separately the methods assessing the return on network services and transmission service charges during the regulatory period. The methods shall be confirmed before the implementation (confirmation decision). According to Section 38 a 2 of the Electricity Market Act, confirmation decision shall be based on the principles prescribed in Section 3, 4 and 6 a of the Electricity Market Act and Act of the European Parliament and European Council 12 (EC No 714/2009). The regulatory methods concerning the pricing of electricity distribution operations and highvoltage electricity distribution operations are based on the pricing principles stated in Section 3 10 and 14 of the Electricity Market Act. According to Section 38 a 2 of the Electricity Market Act, in the confirmation decision Energy Market Authority can determine about: The valuation principles of the capital invested to network operations The methods of measuring reasonable return for the capital invested to network operations methods of determining the result of the system operations and the correction of the income statement and balance sheet required by them Target encouraging improvement of the efficiency of the system operations and the method of determining it, as well as the method to apply the target in pricing 9.2 Regulatory period The length of regulatory period is four years though when ex-ante regulation was introduced, the first regulatory period lasted three years. 1st regulatory period nd regulatory period rd regulatory period Regulation (EC) No 714/2009 of the European Parliament and of the Council of 13 July 2009 on conditions for access to the network for cross-border exchanges in electricity and repealing Regulation (EC) No 1228/

70 The initial guidelines EMV prepared for the third regulatory period ( ) were published January 14th. Separate guidelines for distribution system operators (DSOs) and transmission system operator (TSO) were published. EMV invited comments about the guidelines from the industry and some interest groups. The comments were evaluated and further projects conducted concerning the regulation methods. After inviting and evaluating further comments, finalized guidelines were published on 29th June and sent to DSOs and TSO. Preliminary confirmation decisions (based on finalized guidelines) were sent for comments in September 22nd. After evaluating responses and possibly conducting clarifying projects EMV possibly invites further comments and evaluates them. Confirmation decisions are finalized and sent to DSOs and TSO before Christmas. The third regulatory period starts on January first The differences in regulation between TSO and DSOs The responsibility on the functioning of the electricity transmission system, a task that distribution grid operators do not have, is the main difference in the legal position of TSO and the DSOs. The regulatory model of TSO is based on same principles as the model for DSOs regulation. The models have several similar building blocks Structure of the regulation model and formation of reasonable return and TSO s profit is illustrated in Figure 1. In the following text, the methods differing in the regulation of DSOs and TSO in the third regulatory period are briefly described. A table in chapter list the WACC parameters used in calculation of TSO s reasonable rate of return. 68

71 Figure 1 Structure of the regulation model of TSO in the third regulatory period ( ) Adjustment of the balance sheet and calculation of reasonable return Adjustment of profit and loss account and determining the adjusted return Replacement value of the network The number of network components Unit prices of network components Lifetime and average age of network components = Net present value of electricity network + Other adjusted capital invested in electricity network (book value) Other non-current assets Inventory Short and long-term receivables = Adjusted capital invested in electricity network (including equalisation item) Adjusted equity Adjusted interest bearing debts BCI Inflation adjustment CPI Operating profit (operating loss) + Balance sheet items returned to the operating profit (operating loss) + Net change in connection charges + Network rents + Depreciation according to plan from goodwill - Incentive to Invest + Imputed straight line depreciation - calculated from the + replacement value of network +Depreciation according to plan from network - Incentive to improve quality + DCO reference level - DCO realized - Incentive to improve efficiency - Innovation incentive + Reasonable R&D costs - - Reasonable rate of return (WACC-%) for the adjusted capital invested in electricity network = REASONABLE RATE OF RETURN - = + DEFICIT OR SURPLUS Other adjustments of profit and loss account + Net hedging cost + Cost arising from the financial assets + Imputed taxes = ACTUAL PROFIT (ADJUSTED) Incentive to improve quality Disadvantage caused by outage (DCO) is calculated annually. The method used to calculate DCO is based on connection points of transmission network. In order to calculate the connection point specific DCO, a consumption type specific for each connection point has been defined. Unit prices have been defined for consumption types of different customers in order to calculate the DCO. The incentive to improve quality takes into account connection point specific number of outages, duration of outages and power of outages caused by unexpected outages and autoreclosers. The prices that are used in valuation of connection point-specific outages are based on studies ordered by EMV and Fingrid Oy The formula of DCO calculation is shown in Appendix. The reference level is calculated in similar way as DSOs reference level. The methods are described in chapter and Appendix. EMV considers that the difference between actual disadvantage caused by outages in electricity supply and the reference level of outage costs, which is taken into account in the calculation of actual return on network operations, may correspond with a maximum 13 Pöyry Forest Consulting Oy (2009), Keskeytyksestä aiheutuneen haitan arviointi kemiallisessa metsäteollisuudessa, julkinen raportti, Pöyry Forest Consulting Oy (2009), Keskeytyksestä aiheutuneen haitan arviointi metalli- ja kemianteollisuudessa, julkinen raportti,

72 of two % of the reasonable return calculated for capital invested in the year in question. This applies to both floor and ceiling levels Innovation incentive system The innovation incentive system includes a R&D-expenditure part similar to that in DSOs innovation incentive system. The functioning of system is explained in chapter Incentive to improve efficiency TSO has an efficiency target only on its controllable operative cost (KOPEX). EMV confirms every year of the third regulatory period the starting level of KOPEX. When measuring the starting level EMV uses KOPEX budgeted by TSO in the year in question or the realized KOPEX of five previous years. The network volume is also taken into account. The following equation illustrates the calculation of KOPEX reference level. KOPEX ref, t = [ KOPEX sta, t (1 2,06%)] ± 5% (1) where: KOPEX ref, t = TSO s KOPEX reference level in year t KOPEX sta, t = TSO s KOPEX starting level in year t 2,06% = The general yearly efficiency target for the third regulatory period ± 5% = The error margin of the reference level (This will diminish uncertainty related to the definition of the reference level) Every year the realized KOPEX is subtracted from the KOPEX reference level. The following equation illustrates the calculation of TSO s incentive to improve efficiency TKAN t = kk KOPEX KOPEX [ ref, t real, t ) (2) TKAN, t = TSO s incentive to improve efficiency in year t kk = Incentive coefficient 0,5 (benefit or loss caused by over- or underperforming related to reference level is divided between TSO and customers) KOPEX ref, t = TSOs KOPEX reference level in year t KOPEX sta, t = TSOs realized KOPEX in year t 70

73 Before confirming the starting levels of KOPEX in the incentive to improve efficiency, to be applied in each year of the regulatory period, EMV reserves the right for the TSO to express its views on the confirmed starting levels in a so-called negotiation procedure. The negotiation procedure between EMV and the TSO will take place by 15th of March every year. In the procedure, the actual KOPEX for the previous five years and the budget compiled by the TSO for the KOPEX for the year in question will be discussed. EMV considers that when assessing the reasonableness of pricing of the TSO s network operations, the incentive to improve efficiency in euros to be taken into account in the calculation of the adjusted profit in each year of the third regulatory period may correspond to a maximum of three % of the reasonable return calculated for the capital invested in network operations in the year in question. 9.4 The current regulation in Finland Structure of the regulation model The Finnish electricity distribution regulation model can be described as an ex-ante revenue cap model. EMV uses the regulation model to set the reasonable rate of return but the DSOs set the tariffs themselves. The regulation model is built of many components and Figure 2 illustrates the formation of reasonable return and actual adjusted profit. The reasonable rate of return is subtracted from the actual adjusted profit yearly and the result is surplus (+) or deficit (-). After the regulatory period, 4-year total surplus/deficit is calculated by adding up the yearly surpluses/deficits. DSOs are obligated to compensate the surplus and allowed to compensate the deficit in their price setting on the following regulatory period. Currently there are 85 DSOs in Finland. In the following text, purpose of each component in Figure 2 is briefly explained and the functioning of the regulation model is clarified. 71

74 Figure 2 The principles of the regulation model of DSOs in the third regulatory period ( ) Adjustment of the balance sheet and calculation of reasonable return Adjustment of profit and loss account and determining the adjusted return Replacement value of the distribution network The number of network components Unit prices of network components Lifetime and average age of network components = Net present value of electricity network + Other adjusted capital invested in electricity network (book value) Other non-current assets Inventory Short and long-term receivables = Adjusted capital invested in electricity network (including equalisation item) Adjusted equity Adjusted interest bearing debts Reasonable rate of return (WACC-%) for the adjusted capital invested in electricity network = REASONABLE RATE OF RETURN BCI - Inflation adjustment = CPI + DEFICIT OR SURPLUS Operating profit (operating loss) + Balance sheet items returned to the operating profit (operating loss) + Net change in connection charges + Network rents + Depreciation according to plan from goodwill - Incentive to Invest + Imputed straight line depreciation - calculated from the replacement value of network - Depreciation according to plan from network - Incentive to improvequality + 0,5 DCO reference level - 0,5 DCO realized - Incentive to improveefficiency + Reasonable costs in accordance with the efficiency target - Realized costs in accordance with the efficiency target - Innovation incentive + Reasonable- R&D costs + Reasonable extra costs caused by hourly metering - Other adjustments of profit and loss account + Net hedging cost + Cost arising from the financial assets + Imputed taxes = ACTUAL PROFIT (ADJUSTED) Inflation When calculating the actual adjusted profit, EMV uses inflation adjustment in the incentive to improve quality and in the incentive to improve efficiency. Inflation adjustment is done annually. The net present value of the distribution network is also adjusted. The adjustments will be done on the basis of consumer price index and building cost index (1995=100) so that the average index level for April June 2011 will be used as the index level corresponding to the 2012 value of money, and the average for April June of the previous year to the year in question will be used as the index level corresponding to the value of money in the other years of the regulatory period. The following formula illustrates the change of building cost index for year t BCI where t = BCI BCI t 1 t 2 1 (3) BCI t, BCI t 1 BCI t 2 = change of the building cost index for year t = The average of the building cost index (1995=100) figures for April June in year t-1. = The average of the building cost index (1995=100) figures for April June in year t-2. The change of consumer price index for year t is calculated in the same way. 72

75 9.4.3 Calculating the reasonable return The Reasonable rate of return is calculated by multiplying regulatory asset base (RAB) with WACC - %. RAB consist of equity and interest bearing debts Net present value of the distribution network When determining the value of capital invested in network operations in the calculation methods concerning the return on the electricity network operator s network operations, EMV will not apply the book value of the electricity network because the book value of the network will not necessarily reflect the actual market value of the capital invested in the network due to its previous tax practices. The value of the electricity network will be adjusted in calculation methods concerning the return on network operations to better meet its market value by using its net present value (NPV) instead of its book value. The NPV of the electricity network is determined each year, January first and for each year in the calculation of the value of capital invested in the network operations, the NPV in the situation of the first of January of the year in question, will be used as the value of the network. When determining the capital invested in network operations, the electricity network administrated by the network operator is treated in the same way regardless of whether it is owned or leased by the network operator. In the event that the network operator has leased a network that it administers in whole or part, the leasing arrangement is dissolved in calculations of the return on network operations, whereupon the leased network components are included in the network assets of the network operator and also in the capital invested in network operations. When calculating the replacement value (RV) the unit prices of component specific network components used in the value determination of the network are mainly based on the unit prices presented in the network recommendation of the Finnish Energy Industries (standard unit costs) and, as far as the components used in the above-mentioned value determination of the electricity network are not included in the above-mentioned list of costs, the costs are based on the unit price study concerning the components in question, ordered from Empower Oy by EMV. The standard unit prices are in the value of money of year 2010 and they are adjusted to correspond with the value of money of year 2012 (in the beginning of third regulatory period). Standard values are defined also for the buildings, sites, computer systems etc. invested in network operations. For the other years of the third regulatory period ( ), unit prices are adjusted with inflation according to chapter For justified reasons, it may be possible to take into account enterprise-specific costs due to regional factors instead of the standard unit prices of component-specific network components when calculating the replacement value of the electricity network. NPV is calculated from RV using component-specific unit price and average age data. In order to calculate the average age the DSO is required to find out the true age of the network components. The RV of the whole network is calculated by multiplying all the components with their respective unit prices. In an acquisition, the network acquired is added into the NPV and RV of new owner based on the information on the number of network components and average ages of these components. Respectively, the network is subtracted from the NPV and RV of the seller. 73

76 Calculating the NPV of a network component from the RV, using the average age and lifetime data will be calculated as follows: average age = 1 it NPV it lifetimei RV it (4) NPV it RV it = Net present value of all components i included in network component i in year t, in the 2012 value of money = The combined replacement value of all components included in component i in year t, in the 2012 value of money lifetime i = Lifetime of network component i. The lifetime denotes the period for which a network component is in actual operation before it is replaced (technical financial lifetime). average age = The age information of network component i weighted by its volume information in the beginning of year t The sum of equity and interest bearing debts is multiplied with the WACC-% in order to get the reasonable return in Calculating reasonable rate of return (WACC - %) WACC parameters used in the third regulatory period are mostly based on a consultation report by Deloitte15. Also other documents were used in setting the WACCparameters The parameters used in the third regulatory period are shortly listed in the following table. 15 Deloitte & Touche Oy (2010), Sähköverkkotoiminnan WACC-mallin ja sen parametrien arviointi, EMV s internal memo nro 3 (3nd version)/2011 Justification of WACC parameters Kallunki Juha-Pekka, Sahlström Petri (2010), Lausunto vieraan pääoman kustannukseen sisältyvän riskipreemion ja oman pääoman kustannukseen sisältyvän markkinariskipreemion tarkistamisesta sähköverkkotoiminnalle, Kallunki Juha-Pekka (2011), Lausunto Energiamarkkinaviraston käyttämästä valvontamallista, Market court decisions /2010 (MAO: /10), & (MAO /2006) PricewaterhouseCoopers (2004), Lausunto koskien sähkön jakeluverkkotoiminnan pääoman keskikustannusta Martikainen Teppo (1998), Lausunto Sähkömarkkinakeskukselle jakeluverkkotoimintaan sitoutuneen pääoman kohtuullisesta tuottoasteesta, Kallunki Juha-Pekka (2011), Lausunto Energiamarkkinaviraston käyttämästä valvontamallista, Kuosmanen T. Pursiainen H. Inflaatiokorjaus ja siihen sovellettavat hintaindeksit sähköverkkotoiminnan valvontamallissa

77 WACC parameter DSO TSO Risk free rate (real value) 10 year Finnish government bond yield (average of previous year May) - inflation component 10 year Finnish government bond yield (average of previous year May) - inflation component Inflation component (subtracted from nominal risk free rate) 1,0 % 1,0 % Asset beta 0,4 0,4 Equity beta 0,527 0,844 Market risk premium 5 % 5 % Premium for lack of liquidity Capital structure (Debt / Equity) 0,5 % 0,5 % 30 / / 40 Tax level 26 % 26 % Debt premium 1,0 % 1,0 % Calculating the adjusted profit Connection charges EMV adds the annual net change in transferable and refundable connection charges entered in balance sheet (additions to connection charges in the balance sheet deducted by returned connection charges) in the calculations of reasonableness of pricing as income from the network operations in the year in question Network rents If the network operator has leased either partly or wholly a network under its administration, EMV will dissolve the leasing arrangement in the calculations of the reasonableness of pricing. Leased network assets will be treated in the calculations in the same way as if these network assets were owned by the network operator. Network rents include an earnings - component payable to the owner of the network and a depreciation element corresponding to the ageing of the network. As the network assets administered by the network operator are included in the capital invested in the network operations in calculations of the reasonableness and the reasonable return is defined for the capital, for reasons of consistency, the network rents paid by the network operator and entered in the profit and loss account shall be eliminated in a corresponding manner when calculating the actual return. 75

78 If the network rent paid by the network operator includes, for example, operating and maintenance costs for the leased network, and the network operator chooses not to eliminate these costs when calculating the actual return, the network operator must submit to the EMV a separate account of the proportion of these costs in the network rent Depreciation according to plan from goodwill With respect to depreciation of other long-term assets of the network operations, depreciation used in the bookkeeping will be applied when calculating the network operator s actual adjusted return on network operations. On calculations of adjusted actual profit, the depreciation from goodwill in DSO s adjusted balance sheet is added into the operating profit. Correspondingly goodwill not allocated will be eliminated from the balance sheet when determining adjusted capital invested in network operations Incentive to invest The purpose of the incentive to invest is to encourage adequate investments and development of the network. Incentive to invest includes two parts. The first is depreciation method applied to the calculation of adjusted profit. The second is monitoring of adequate investment level of the DSO Depreciation method of the incentive to invest Depreciation method of the incentive to invest includes depreciation based on imputed straight-line depreciation determined from the network replacement value and the depreciation according to plan from the electricity network in the network operator s bookkeeping. The imputed straight-line depreciation will be determined annually based on the network replacement value. The determination of the replacement value is described in chapter of this document. The imputed straight-line depreciation will be determined by network component in the beginning of each year. With respect to one network component i, the imputed straight-line depreciation in the year t (SLDt,i) will be calculated as follows: SLD t, i = RV t, i lifetime i (5) For the entire network, the imputed straight-line depreciation will be calculated as a sum of straight-line depreciations determined for individual components. The imputed straight-line depreciation for the network in the year t (SLDt) is calculated as follows: SLD t = n i= 1 RVt, i lifetime i (6) In the above formula SLD t, i = Imputed straight-line depreciation of component i in the year t SLD t = Imputed straight-line depreciation of the network in the year t 76

79 RV t, i = Replacement value of a component or component group i in the year t lifetime i = Lifetime of network component i. The lifetime denotes the period for which a network component is in actual operation before it is replaced (technical-financial lifetime). If the network operator has leased either partly or wholly a network under its administration, EMV will dissolve the leasing arrangement in the calculations of the network operator s return on network operations. Leased network assets will be treated in the calculations in the same way as if the said network assets were owned by the network operator. In this case network depreciation also includes straight-line depreciation calculated on the leased network assets as described above Monitoring the adequate investment level The purpose of straight line depreciation calculated from the replacement value is to ensure that DSOs have adequate income level for necessary replacement investments. The level of replacement investments vary yearly and replacement investment surplus or deficit may occur. This surplus or deficit created in the 3rd regulatory period can be balanced in the following regulatory periods. The level of replacement investment and imputed straight line depreciation is monitored by EMV in the 3rd regulatory period. If imputed straight line depreciation exceeds the actual replacement investment, so called replacement investment deficit occurs. In the opposite case, so called replacement investment surplus occurs. Figure 2 illustrates the effect of imputed straight line depreciation in the calculation of adjusted profit of DSO Incentive to improve quality Calculating the disadvantage caused by outages in electricity supply In the annual calculation of the disadvantage caused by outages (DCO) in the electricity supply of network operators, the planned and unplanned outages in long outages and the high-speed and time-delayed autoreclosers in short outages will be taken account. In the calculation, the key figures on the number and duration of outages, describing outages specific for particular network operators and reported to the Energy Market Authority by the network operators, will be used as the outage data. The prices used for various outages are based on the results of a study on the disadvantage of electricity supply outages, which was carried out in The prices of outages obtained as the result of the study have been revised so that, combined with the key figures on outages, gathered annually from the distribution network operators, the cost of electricity supply outages will describe the cost for the customers as accurately as possible. A description of the way the results of the original 24 Teknillinen korkeakoulu, Tampereen teknillinen yliopisto / Silvast Antti, Heine Pirjo, Lehtonen Matti, Kivikko Kimmo, Mäkinen Antti, Järventausta Pertti (2005), Sähkönjakelun keskeytyksistä aiheutuva haitta, joulukuu

80 study have been revised is presented in the end reports of studies commissioned by the Energy Market Authority The calculation of actual disadvantage to the customer caused by electricity supply outages is illustrated in Appendix Determining the reference level of disadvantages caused by outages in electricity supply and calculating the adjusted return In the third regulatory period, the arithmetic average of the network operator s actual outage costs in adjusted to the value of money for the year in question will be used as the reference value for the network operator s outage costs in various years of the third regulatory period. This way the reference level represents the normal outage level. When calculating the reference value for outage costs to be applied in each year of the regulatory period, the outage costs in will still be adjusted so that they will correspond to the outage costs calculated with the annual energy transferred to the customer during the year under review. This way, the impact of fluctuation in annual energy transferred to the customers on the reference value of outage costs will be eliminated. The DCO in the network operator s electricity supply when calculating the adjusted profit will be taken into account through the incentive to improve quality method of the regulatory model Also the major deviations in outages are taken into account by setting a limit value (socalled floor and ceiling levels) for the difference between the actual DCO in electricity supply and the reference level of DCO to be taken into account in different years of the regulation model. The formula for calculation of the reference level is illustrated in Appendix. A half of the difference between actual disadvantage caused by outages in electricity supply and the reference level of outage costs, which is taken into account in the calculation of actual return on network operations, may correspond with a maximum of 20 % of the reasonable return calculated for capital invested after tax in the year in question. This applies to both floor and ceiling levels Innovation incentive system Innovation incentive system includes two parts: the expenditure caused by research and development (R&D) activities and the costs for DSOs caused by hourly metering system (place of electricity use with max 63A main fuse). According to law 27, by the end of 2013 hourly metering has to cover 80 % of distribution network operators places of electricity use (max 63A). Based on a study of Pöyry 25 Honkapuro Samuli, Tahvanainen Kaisa, Viljainen Satu, Lassila Jukka, Partanen Jarmo, Kivikko Kimmo, Mäkinen Antti, Järventausta Pertti (2006): DEA-mallilla suoritettavan tehokkuusmittauksen kehittäminen (Development efficiency measurement performed with the DEA model). Lappeenranta University of Technology Honkapuro Samuli, Tahvanainen Kaisa, Viljainen Satu, Partanen Jarmo, Mäkinen Antti, Verho Pekka, Järventausta Pertti (2007): Keskeytystunnuslukujen referenssiarvojen määrittäminen (Determining the reference values in outage key figures). Lappeenranta University of Technology and Tampere University of Technology

81 Management Consulting Oy 28 The Energy Market Authority has measured a reasonable incentive level of yearly operational expenditure which is 5 for every place of electricity use (max 63A) under the hourly metering system. The monetary value of incentive is calculated by multiplying 5 with the amount of places of electricity use (max 63A main fuse) owned by DSO. The sum is subtracted from the operating profit (loss) when calculating the actual adjusted profit. The R&D expenditure that EMV approves as part of the innovation incentive system can be maximum 0,5 % of the annual turnover of DSO. This amount can be subtracted yearly from the operating profit (loss) when calculating the actual adjusted profit Other adjustments of profit and loss account This item includes cost arising from the financial assets necessary to safeguard network operations and hedging costs, which are subtracted from the adjusted operating profit. After that, imputed corporation tax is subtracted to get the adjusted actual return of DSO. 9.5 Applied benchmarking method Incentive to improve efficiency The incentive to improve efficiency contains an efficiency target that includes a company-specific efficiency figure (estimated for each DSO) And a general efficiency target (same for each DSO). The method used in estimating the company specific efficiency target is a semiparametric method called StoNED (Stochastic Nonsmooth Envelopment of Data) 29,30 Applying the StoNED-method to Finnish electricity distribution regulation was proposed in consultation reports conducted by EMV in ,32,33 EMV has applied the general efficiency target since the first regulatory period. The yearly target of 2,06%, based on general productivity and technological progress was initially estimated in The figure was estimated with updated data in The most 27 Valtioneuvoston asetus sähköntoimitusten selvityksestä ja mittauksesta (VNa 66/2009) 28 Pöyry Management Consulting Oy / Antila Heli, Huumo Mikko, Leinonen Juha, Aaltonen Päivi, Mäkinen Sami, Porri Ossi, Saarnio Kaj, Tefke Joni (2010), Tuntimittauksesta aiheutuvien operatiivisten kustannusten huomioiminen valvontamallissa, Kuosmanen Timo, Kortelainen Mika, Stochastic non-smooth envelopment of data: semi-parametric frontier estimation subject to shape constraints, Journal of Productivity Analysis, Volume 35 Issue 2, , DOI: /s , Direct link: 30 StoNED method s website: 31 Sigma-Hat Economics Oy / Kuosmanen Timo, Kortelainen Mika, Kultti Klaus, Pursiainen Heikki, Saastamoinen Antti, Sipiläinen Timo (2010), Sähköverkkotoiminnan kustannustehokkuuden estimointi StoNED-menetelmällä, Sigma-Hat Economics Oy / Kuosmanen Timo (2010), Lisäselvitys kustannusmuuttujan määrittelyn vaikutuksista tehokkuusestimoinnin tuloksiin ja tehostamistavoitteisiin, Lisäselvitys selvitykselle: Sähköverkkotoiminnan kustannustehokkuuden estimointi StoNED-menetelmällä, Kuosmanen Timo (2010), Cost efficiency analysis of electricity distribution networks: Application of the StoNED method in the Finnish regulatory model, Direct link: 34 Korhonen Pekka, Syrjänen Mikko (2003): Lausunto sähköverkkotoiminnalle asetettavasta yleisestä tehostamisvaatimuksesta. Helsingin kauppakorkeakoulu, Gaia Consulting Oy / Syrjänen Mikko (2007), Lausunto tuottavuuskehityksen huomioivasta alan yleisestä tehostamistavoitteesta,

82 recent study suggested a figure very close to 2,06% 36 so EMV has found no need to change the size of general efficiency target Model specification The recommended model specification is the following: TOTEX = C y ) exp( δ z i + ε ) = C( y ) exp( δ z i + u + v ) i ( i i i i i (7) where TOTEX i = Total cost of DSO i (average of years ) C = Frontier cost function y i = The output vector of DSO i (average of years ) δ = A parameter characterizing the effect of underground cabling (1-70 kv) on total cost of DSO i z i = The proportion of underground cables (1-70 kv) in the total length of network i (average of years ) ε i = u i + vi = A combined error term u i is v i is u i v i = A random variable representing cost inefficiency of DSO i (average of years ) = A stochastic noise term that captures the effects of measurement errors, omitted variables and other disturbances to the otherwise stable cost function Input In the estimation of company specific efficiency target EMV uses data on DSOs input, outputs and contextual variables from years Calculation of the reference level of the input variable is illustrated in equation 8. TOTEX where i, , avg ( KOPEX 6 CPI ) CPI 2010 i, t + 0,5 DCOi, t 2010 = 2005 = t t (8) TOTEX i, , avg = Average (years ) of total costs of DSO i indexed into year 2010 value of money. KOPEX i, t = The realized controllable operative costs 37 of DSO i in the year t 36 Sigma-Hat Economics Oy / Kuosmanen Timo, Kortelainen Mika, Kultti Klaus, Pursiainen Heikki, Saastamoinen Antti, Sipiläinen Timo (2010), Sähköverkkotoiminnan kustannustehokkuuden estimointi StoNED-menetelmällä,

83 DCO i, t = Actual disadvantage caused by electricity supply outages to the customers of the DSO i in the year t CPI 2010 CPI t = The CPI index point level in 2010 = The CPI index point level in year t Outputs Outputs used in the model are y1 = The amount of energy transmission (GWh) y2 = Length of network (km) y3 = Number of customers The amount of energy distribution y1 captures the direct (variable) output of the distribution activity. The amount of energy distribution is weighted with the average national distribution prices of years (0,4 kv, 1-70 kv and 110 kv). Using fixed weights ensures that comparison between different years will not be affected by inflation or changes in the electricity. Outputs y2 and y3 represent the potential output or the capacity. These variables capture the fixed cost of maintaining a sufficient capacity to serve their designated network area irrespective of the actual consumption of power Contextual (operating environment) variable While the output variables y2 and y3 can draw a distinction between urban versus rural networks, the networks located in large cities have very similar output structure as those located in suburbs or small towns. To better capture the differences in the output structures and operating environments of the urban versus suburban networks, a contextual variable z has been placed into the model. The proportion of medium voltage (1-70 kv) underground cables is used as z-variable. The proportion is calculated based on the total network length (average of years ) The reasonable (allowed) total cost level (STOTEX) The company specific efficiency figure (%) can simply be calculated as shown in equation 9. TL i where TOTEX = C i i (9) TL i TOTEX i C i = Efficiency figure of DSO i = Total cost of DSO i realized (average of years ) = The value of cost frontier C estimated with the outputs of DSO i (average of years ) 37 See Appendix for the list of KOPEX 81

84 In order to get the company-specific efficiency target, the efficiency figure of DSO i and general efficiency target are combined. The transition period is taken into account in the calculation. The transition period of eight years has been set because differences in efficiency between DSOs appear to be wide and this way the cost adjustment into efficient level can be softened. Equation 10 shows the calculation of company-specific efficiency target. Xi is the company specific efficiency target and 2,06% is the general efficiency target. X 1 8 i TL (1 2,06%) = i (10) Equation 11 illustrates the calculation of reasonable (allowed) total cost level (STOTEX) STOTEX = C 2011 ( y i, t i, t = C 2019 ( y ) exp( δz i, t i, t ) exp( δz CPI ) CPI i, t t CPI ) CPI t (1 2,06%) (1 X 8 i ) t 2019 t ( 1 X ) 2019 i (11) where STOTEX i, t = Reasonable total cost level of DSO i in year t CPI t 1 CPI 2009 X i = Inflation index in year t-1 = Inflation index in year 2009 = Company specific-efficiency target of DSO i 2019 C = Cost frontier of DSO s in year 2019, in year 2009 price level y i, t = Output vector of DSO i in year t z i, t = Proportion of underground cables (1-70kV) DSO i in year t Growth in the level of outputs and proportion of underground cables is taken into account by calculating yearly the value of total cost C(y). The calculation of yearly reasonable cost level (STOTEX) can be done with an Excel application found in the EMV website. STOTEX is yearly subtracted from the realized total cost in the calculation of adjusted profit (see Figure 2). 82

85 9.6 Evaluation of regulation and issues in future regulation On the request of the Energy Market Authority, Lappeenranta University of Technology conducted an evaluation on the Finnish electricity distribution model. The objective of the project was to evaluate functionality and steering effects of the regulation model. Based on these the project group should assess development needs of the model. The report was part of the Roadmap 2020 project. The Energy Market Authority started Roadmap 2020 project to find out megatrends of near future in the electricity distribution field and develop the regulation model The project group studied regulation model by evaluating data of grid operators. The report included also a survey and network operators were interviewed about their opinion on the different parts of the regulation model. The incentivizing effect of the regulatory model was analysed. The report was conducted in 2010 and it focused on the second regulatory period. Overall the regulation model was found to function all well. A few development ideas were suggested and the Energy Market Authority has taken some of these suggestions into account when preparing the model for the third regulatory period. In 2008 EMV started a Roadmap 2020 project. The objective of this project was to find out the megatrends in electricity distribution in the near future, develop the regulation model and define a strategy for regulation for 2020 and onwards. There were four megatrends in network industry identified in the project: 1. Significant need for replacement investments Network lifetimes are coming to end The arrival of smart meters and smart networks DSOs will concentrate on their core competence and strategy (more outsourcing) 2. Society s increased reliance on electricity Interruptions will not be accepted Exceptional storms will happen more often 3. Emission free electricity production increases Increased energy production using wind farms and bio-fuels Two-way distribution and micro production units More nuclear power in networks 4. Integrated regulation Detailed steering and regulation at EU level (Acer) It is possible that above issues have to be concerned in the future development of the regulation model. 83

86 9.7 Appendix Controllable operating costs of DSO (KOPEX) KOPEX = Materials, accessories and energy purchases + Increase or decrease in stocks + Staff costs + Rents + Other external services + Internal costs (with respect to 2005 and 2006) + Other costs + Standard compensations paid (if not included in other costs) - Production for own use Calculation of Disadvantage Caused by Outage for TSO TSO conducts annually the calculations about the disadvantage caused by connection point-specific outages. The following formula is used in the calculation. DCO where = n CPI k 1 [ ] ( A + i Bi Ti Pi K i, va K i, vp i= 1 CPI 2009 t, k ) DCO t, k = Actual disadvantage caused by electricity supply outages to the customers of the network operator in year t in the value of money of year k. n = The number of unexpected outages in year t. A i B i T i P i K va = A wattage-coefficient of unexpected outage i which depends on the consumption type. = An energy-coefficient of unexpected outage i which depends on the consumption type. = duration of unexpected outage i, hours. = The wattage of connection point when the unexpected outage i starts, kw. = A season-coefficient of the starting time of unexpected outage i. 84

87 K vp = A time of the day -coefficient of the starting time of unexpected outage i. CPI k 1 = Consumer price index in year k-1 CPI 2009 = Consumer price index in year 2009 Calculation of Disadvantage Caused by Outage for DSOs In the assessment of the electricity network operator s return on network operations, the actual disadvantage to the customer caused by electricity supply outages (hereinafter also the same as the outage cost) in the year i (DCOt,k) in the value of money in the year k will be calculated with the following calculation formula: DCO t, k OT OT un exp, t h h E, un exp + ON ON un exp, t h W, un exp + W t = plan, t E, plan + plan, t W, plan + T 1 t TDAt htda + HSAt hhsa h ( + CPI ) k where DCO t, k = Actual disadvantage (outage cost) caused by electricity supply outages to the customers of the network operator in the year t in the value of money of the year k. OTun exp, t = Customer s average annual outage time weighted by annual energies, caused by unexpected outages in the 1 70 kv network in the year t, hour h E,un exp = Price of disadvantage caused by unexpected outages to the customer, EUR/kWh, in the 2005 value of money. ON un exp, t = Customer s average annual number of outages weighted by annual energies, caused by unexpected outages in the 1 70 kv network in the year t, number. h W,un exp = Price of disadvantage caused by unexpected outages to the customer, EUR/kW, in the 2005 value of money. OT plan, t = Customer s average annual outage time weighted by annual energies, caused by planned outages in the 1 70 kv network in the year t, hour h E, plan = Price of disadvantage caused by planned outages to the customer, EUR/kWh, in the 2005 value of money. 85

88 ON plan, t = Customer s average annual number of outages weighted by annual energies, caused by planned outages in the 1 70 kv network in the year t, number. h W, plan = Price of disadvantage caused by planned outages to the customer, EUR/kW, in the 2005 value of money. TDA t h TDA HSA t h HSA W t T t CPI k = Customer s average annual outage number weighted by annual energies, caused by time-delayed autoreclosers in the 1 70 kv network in the year t, number = Price of disadvantage caused by time-delayed autoreclosers to the customer, EUR/kW, in the 2005 value of money. = Customer s average annual outage number weighted by annual energies, caused by high-speed autoreclosers in the 1 70 kv network in the year t, number = Price of disadvantage caused by high-speed autoreclosers to the customer, EUR/kW, in the 2005 value of money. = The amount of energy transferred to the users from the network operator s network with a voltage of 0,4 kv and 1 70 kv in the year t, kwh. = Number of hours in the year t. = Change in the consumer price index for the year k. The change in the building cost index in the above formula for calculating outage costs for the year k will be calculated with the following formula: CPI k CPI = CPI k where CPI k CPI k = Change in the consumer price index for the year k. = The average of the consumer price index (1995=100) index figures for April June in the year k. The reference value (DCOref,k) of the disadvantage caused by the network operator s outage costs for the year k during the regulatory period of will be calculated using the following formula: 86

89 DCO where ref, k 2010 k DCO, t k = 2005 W = t t 6 W DCO ref, k = The reference value of disadvantage caused by outages in the network operator s electricity supply (outage costs) for the year k. DCO t, k = Actual disadvantage caused by electricity supply outages to the customers of the network operator in the year t in the value of money of the year k. W k W t = The amount of energy transferred to the users from the network operator s network with a voltage of 0,4 kv and 1 70 kv in the year k, kwh. = The amount of energy transferred to the users from the network operator s network with a voltage of 0,4 kv and 1 70 kv in the year t, kwh. 9.8 References Deloitte & Touche Oy (2010), Sähköverkkotoiminnan WACC-mallin ja sen parametrien arviointi, EMV s internal memo nro 3 (3nd version)/2011 Justification of WACC parameters (In Finnish) Gaia Consulting Oy / Syrjänen Mikko (2007), Lausunto tuottavuuskehityksen huomioivasta alan yleisestä tehostamistavoitteesta, Honkapuro Samuli, Tahvanainen Kaisa, Viljainen Satu, Lassila Jukka, Partanen Jarmo, Kivikko Kimmo, Mäkinen Antti, Järventausta Pertti (2006): DEA-mallilla suoritettavan tehokkuusmittauksen kehittäminen (Development efficiency measurement performed with the DEA model). Lappeenranta University of Technology Honkapuro Samuli, Tahvanainen Kaisa, Viljainen Satu, Partanen Jarmo, Mäkinen Antti, Verho Pekka, Järventausta Pertti (2007): Keskeytystunnuslukujen referenssiarvojen määrittäminen (Determining the reference values in outage key figures). Lappeenranta University of Technology and Tampere University of Technology Kallunki Juha-Pekka, Sahlström Petri (2010), Lausunto vieraan pääoman kustannukseen sisältyvän riskipreemion ja oman pääoman kustannukseen sisältyvän markkinariskipreemion tarkistamisesta sähköverkkotoiminnalle, Kallunki Juha-Pekka (2011), Lausunto Energiamarkkinaviraston käyttämästä valvontamallista, Korhonen Pekka, Syrjänen Mikko (2003): Lausunto sähköverkkotoiminnalle asetettavasta yleisestä tehostamisvaatimuksesta. Helsingin kauppakorkeakoulu,

90 Kuosmanen T. Pursiainen H. Inflaatiokorjaus ja siihen sovellettavat hintaindeksit sähköverkkotoiminnan valvontamallissa Kuosmanen Timo (2010), Cost efficiency analysis of electricity distribution networks: Application of the StoNED method in the Finnish regulatory model, Direct link: Kuosmanen Timo, Kortelainen Mika, Stochastic non-smooth envelopment of data: semiparametric frontier estimation subject to shape constraints, Journal of Productivity Analysis, Volume 35 Issue 2, , DOI: /s , Direct link: Market court decisions /2010 (MAO: /10), & (MAO /2006) Martikainen Teppo (1998), Lausunto Sähkömarkkinakeskukselle jakeluverkkotoimintaan sitoutuneen pääoman kohtuullisesta tuottoasteesta, PricewaterhouseCoopers (2004), Lausunto koskien sähkön jakeluverkkotoiminnan pääoman keskikustannusta Pöyry Forest Consulting Oy (2009), Keskeytyksestä aiheutuneen haitan arviointi kemiallisessa metsäteollisuudessa, julkinen raportti, Pöyry Forest Consulting Oy (2009), Keskeytyksestä aiheutuneen haitan arviointi metalli- ja kemianteollisuudessa, julkinen raportti, Pöyry Management Consulting Oy / Antila Heli, Huumo Mikko, Leinonen Juha, Aaltonen Päivi, Mäkinen Sami, Porri Ossi, Saarnio Kaj, Tefke Joni (2010), Tuntimittauksesta aiheutuvien operatiivisten kustannusten huomioiminen valvontamallissa, Regulation (EC) No 714/2009 of the European Parliament and of the Council of 13 July 2009 on conditions for access to the network for cross-border exchanges in electricity and repealing Regulation (EC) No 1228/2003 Sigma-Hat Economics Oy / Kuosmanen Timo, Kortelainen Mika, Kultti Klaus, Pursiainen Heikki, Saastamoinen Antti, Sipiläinen Timo (2010), Sähköverkkotoiminnan kustannustehokkuuden estimointi StoNED-menetelmällä, Sigma-Hat Economics Oy / Kuosmanen Timo (2010), Lisäselvitys kustannusmuuttujan määrittelyn vaikutuksista tehokkuusestimoinnin tuloksiin ja tehostamistavoitteisiin, Lisäselvitys selvitykselle: Sähköverkkotoiminnan kustannustehokkuuden estimointi StoNED-menetelmällä, StoNED method s website: Teknillinen korkeakoulu, Tampereen teknillinen yliopisto / Silvast Antti, Heine Pirjo, Lehtonen Matti, Kivikko Kimmo, Mäkinen Antti, Järventausta Pertti (2005), Sähkönjakelun keskeytyksistä aiheutuva haitta, joulukuu 2005 Valtioneuvoston asetus sähköntoimitusten selvityksestä ja mittauksesta (VNa 66/2009) 88

91 10 Country report: Norway 10.1 The objective of the regulation Introduction The structure of the management of energy and water resources Stortinget (the parliament) lays down the political framework for the management of energy and water resources. Based on this framework, the Ministry of Petroleum and Energy (MPE) has the principal responsibility to achieve a coordinated and integrated energy policy. The Norwegian Water Resources and Energy Directorate (NVE) is a subordinate agency of MPE. MPE is appellate body for decisions made by NVE The Norwegian Water Resources and Energy Directorate (NVE) NVE was assigned the role of electricity regulator when the Norwegian Energy Act entered into force on 1 st of January The act authorizes the framework of regulations and licenses necessary to establish and regulate an efficient power market, with free choice of supplier and regulated access to the networks, and the issuing of regulations concerning the rights and obligations of the various actors in the market. NVE is authorized to monitor compliance with, and make decisions according to, the Energy Act and regulations laid down in accordance with the Act. NVE is also authorized to issue regulations in main areas important to securing an efficient electricity market, such as network regulation and tariffs, quality of supply, metering and settlement, billing, supplier switching, neutrality and non-discrimination, and, finally, the obligations and powers of the transmission system operator (TSO) Framework for the economic regulation of network companies The regulatory framework for economic regulation of the network companies is formulated based on objectives given in the Energy Act and regulations laid down in accordance with the act. The key objectives are: Energy Act Regulation 4-1: ( ) to secure a socially rational energy sectors and network through enabling an effective energy market and an effective management, utilization and development of the electricity network ( ) Energy Act Regulation 4-4b: NVE shall annually determine revenue caps for each company. Their income shall over time cover the costs of management and depreciation of the network, as well as give a reasonable rate of return on invested capital given effective management, utilization and development of the network. Regulation governing financial and technical reporting, revenue caps for network operations and transmission tariffs (Reg. No. 302 of 11 March 1999): ( ) shall lay the basis for an efficient power market and controls on network operation as a natural monopoly. ( ) ensure that power is transmitted at the correct price and quality of supply and that the network is utilized and developed safely and in a way that efficiently promotes the interests that are affected. 89

92 Direct regulation Economic and direct instruments must be combined in the regulation of power networks in order to reach an optimal adaptation and at the same time avoid unwanted external effects. This is why the power networks are subject to many sets of rules in acts, regulations and conditions for licensing which govern their duties and rights. These regulations lead the power networks activity, and shall ensure that the companies invest and maintain the network, that safety- and emergency preparedness concerns are addressed, that the quality of supply is sustained adequately, that the security of supply is maintained in demanding situations, that the electrical network are adequately robust, and that investments with large environmental disadvantages are not carried through if society s benefit from the investments are lower than the social costs. NVE expects the network companies to follow the rules, and that socio-economically rational actions are carried out according to these. Compliance with the direct regulations is controlled through comprehensive monitoring activity each year Tariffs NVE does not determine each network company`s electricity distribution tariffs, but sets an upper limit for the revenue the company can recover from its customers. In addition, NVE has made a set of rules that decide the structure of the tariffs. According to current regulation, the tariff structure shall vary between customers connected to the central and regional grid (transmission and high-voltage distribution), and customers connected to the distribution grid. In the central and regional grid, the tariff must consist of an energy component (NOK/kWh) and a power component (NOK/kW). The energy component is equal to the Nord Pool system price (NOK/MWh) multiplied with the marginal loss rate at the point of connection. The marginal loss rates in central and regional grid are updated every week, and there are different rates during daytime and night time/weekends. The power component is settled based on the customer s average total consumption at peak hour over the last five years per connection point. Power producers pay a residual tariff based on a ten-year average of the production in MWh, in addition to a marginal loss component. The rate of the residual tariff is set for the central grid and is normative for the lower grid levels as well, so that the residual tariff cost is the same for all production independent on the grid level it is connected to. The level of the residual tariff may vary within a range of 0 to 1,2 /MWh set by Commission regulation 774/2010. Tariffs to household customers must consist of a fixed component (NOK/year) and an energy component (NOK/kWh). The fixed component is a fixed annual amount and shall at minimum cover customerspecific costs. These includes costs related to metering, settlement, invoicing etc. The energy component depends on consumption and shall at minimum cover costs of marginal loss (the loss that occurs when one extra kilowatt-hour is taken out, at a given load) in the network. In addition, the tariffs cover the remaining fixed costs in the network. Many network companies choose to divide their customers in groups, charging different tariffs from different customer groups. For example, it is not unusual to have different tariffs for household, industry and cottages. 90

93 Electricity distribution tariffs usually vary from company to company. This is mainly due to the difference in companies` costs related to owning and operating the network and the difference in the companies` scale between the fixed and the energy component. To contribute to leveling tariffs between distribution companies, there is an annual grant through the government budget given to reduce the distribution tariffs in areas with highest costs per kwh distributed. The amount of the total grant differs from year to year, and has been in the range of million NOK. The government has proposed to increase it to 120 MNOK in NVE transfers the annual amount to the distribution companies with highest costs per unit, at terms given by the MPE. The transferred amount is used to lower the companies distribution tariffs Regulation period The economic regulation of the network companies is executed by annual determination of revenue caps for each company. The current revenue cap regulation entered into force as of 1 st of January According to the Energy Act Regulation, the main principles for calculation of revenue cap shall be re-evaluated periodically. Each period shall last a minimum of 5 years. NVE is not planning any major changes in the principles for calculation when the minimum period of 5 years expire at the end of However, the model for calculation of the cost norm has been under constant evaluation since the model came into force in 2007, and the cost norm model has been revised several times. The purpose of these revisions is to ensure that the model for benchmarking is taking into account new knowledge on all relevant data regarding the companies operational conditions, such as differences in geographical, climate and structural conditions Current regulation Grid structure The electricity grid in Norway is divided into three levels; distribution grid, regional grid and central grid. The distribution grid is defined as the networks with voltages up to and including 22 kv. The regional grid is a regional transmission system between the distribution and central grid, with voltages from 33 kv up to and including 132 kv. The central grid includes all voltages above 132kV, still mainly a voltage of 300kV and 420 kv, but also some 132 kv lines are included. The central grid is a national transmission system. In 2011, Norway has one transmission system operator (TSO); Statnett SF. The TSO owns most of the central grid (approximately 90 % in 2011). Further, there are 158 network companies, whereas 142 have distribution grid and 91 have regional grid. Several companies have both distribution and regional grid, and a few also owns some of the central grid. In the regulation; distribution grid, regional grid and the central grid not owned by the TSO (appr. 10 %), is regulated using the same model. To make it comparable with other countries regulation, this model is called the DSO model in this report. 91

94 The DSO model NVE decides a yearly revenue cap for all network companies. The revenue cap is calculated according to the formula: RC = t * 0.4Ct Ct (1) RC t is the revenue cap in year t. C t is the cost base for each network company, based on costs from year t-2. C t * is the cost norm for the company, which is the result of a benchmarking analysis of the companies, also based on data from year t-2. NVE use Data Envelopment Analysis (DEA) to benchmark the companies costs (described in chapter 11.6). The cost base is calculated according to the formula: CPI C t ( OM 2 + CENS 2 ) + PLt 2 Pt + DEPt 2 + RABt 2 WACCt t = t t CPIt 2 (2) OM is the operation and maintenance cost for the company and CENS is the company s costs of energy not supplied. These costs are inflated by CPI to reflect the cost level in the revenue cap year t. The cost of power losses is calculated by multiplying actual power loss (PL) with a reference price of power (P). The reference price is the relevant weighted average area spot price in year t. DEP is annual depreciations and RAB is the regulatory asset base. The WACC is the weighted average cost of capital, and shall reflect a reasonable return on capital in year t. NVE notifies the expected revenue cap before the start of the year, based on estimates for WACC, CPI and the reference price on power related to power losses. These parameters are not known until the revenue cap year ends. When the parameters are known, NVE decides the final revenue cap for the year based on the actual values. The main intension of notification of the revenue cap is to give the companies an approximate figure to base their tariff decisions on for the revenue cap year. The revenue caps are calculated based on expected total costs, using CPI adjusted costs from year t-2 (e.g. the revenue cap for 2011 is based on reported data from 2009). If there is a deviation between the expected total costs and the actual costs of the network sector in a year, a correction of the deviation is included in the revenue cap calculation two years later (e.g. a deviation between expected and actual costs for 2009 will be corrected in the revenue cap for 2011). The deviation is distributed to the companies using their share of the sectors total RAB. The companies set their tariffs based on their allowed revenue. The allowed revenue (AR) for each company is calculated according to the formula: [ DEP DEP ) + ( RAB RAB WACC ] AR t = RCt + PTt + TCt CENSt + ( t t 2 t t 2 ) t (3) PT is the property tax. TC is tariff costs to other regulated grids. These costs are not included in the revenue cap itself, but added to the allowed revenue. The CENS costs subtracted in the AR-formula are actual CENS costs, without time lag e.g. in the AR for 2011, the subtracted CENS cost are related to outages that took place in The expression in [ ] is included to remove the two year time lag on capital costs related to 92

95 any changes in these. The implication of this element is that the companies can calculate the capital costs related to investments into their allowed revenue in the actual year the investments are made. The allowed revenue constitutes a basis for each company s calculation of tariffs, and the revenue compliance is subject to regulatory control. Excess or deficit revenue for a given year is calculated as the difference between actual collected revenues (CR) in a year t and allowed revenue for year t: Excess / deficit t = CR t AR t (4) The compliance of excess/deficit-revenues is regulated by NVE, through the Regulation on the economic regulation of the power networks. Every year, when all relevant data regarding revenue cap, allowed revenues and actual collected revenues are known, NVE decides an excess/deficit revenue balance for each company. This decision is made approximately one year after the revenue cap is set, when the companies have reported their actual costs in the revenue cap-year. The excess/deficit revenue balance is to be adjusted towards zero over time, through tariff changes. Excess revenues must be reimbursed to the customers, while deficit revenues may be recovered. The changes in tariffs due to excess/deficit revenue balance are not to be done over night, the effect of this method is smoother tariffs and avoidance of substantial changes in tariffs from year to year Total costs All economic, and some technical, data are collected annually through NVEs report system called erapp. Total costs that enter into the regulation, consist of the companies operation and maintenance costs, capital costs, costs related to network losses and CENS. The costs are mainly calculated by the same principles for both DSOs and TSO (see chapter 11.5 for illustration of differences) Operation and maintenance costs The operation and maintenance costs mainly comprise of personnel- and material costs related to own production of services and purchase of external services. The following cost units are included: purchases of transmission and system services, purchases of goods and material, salaries and other personnel costs incl. pension costs, other operating costs, bad debts and internally priced services. A network company can charge a customer for costs related to existing network when the customer asks for quality or services that are usually not expected to be delivered as a network service. Revenues following these kinds of services can be kept out of actual revenue, but are also subtracted from the cost base Inflation NVE use the annual average Consumer Price Index (CPI) to inflate the O&M costs and CENS from two years back to reflect the price level in the revenue cap year t. The CPI is collected from Statistics Norway ( Capital costs The capital costs that enter into the regulated cost base are annual depreciations and return on the regulatory asset base (RAB). The RAB is based on historical book values 93

96 for networks and other fixed assets of the network companies. All accumulated depreciations and write-downs are subtracted from the historical cost. Networks under construction are excluded from the RAB. When calculating the RAB, NVE use net book values per RAB The RAB consists of the following fixed assets: central-, regional- and distribution networks, sites, buildings, customer-specific equipment, means of transport, fixtures, tools and computer equipment, and other fixed assets used in the network business. Leased assets are not included in the RAB. However, the leasing costs of network assets (lines, cables, transformers, switches) which is leased from another regulated network company is allowed to include in the tariffs costs related to other regulated grids (TC). Hence, it will be included into the allowed revenue (AR). Assets that are financed by third parties as investment contributions, are not included in the RAB, however, O&M costs related to these are included in the company s O&M costs. To reflect the working capital, NVE add 1 % of the total book value per to the RAB. The RAB is then multiplied with the WACC to calculate a reasonable return on the regulatory asset base WACC The Weighted Average Cost of Capital (WACC) is given by the following formula: WACC post tax [ Rf ( 1 t) + e MP] + 0.6( Rf + Pd ) ( t) = 0.4 β 1 (5) Where the variables are as follows: Nominal risk free rate (Rf): Annual average yield of government bond with 5 years maturity Tax rate (t): 28 % Market premium (MP): 4 % Asset beta: 0,35 (equity beta βe: 0,875) Debt/equity share: 60/40 Debt premium (Pd): 0,75 % A CAPM-model was used to assess the cost of the companies equity. The gearing ratio is 60 %. An average 5-years maturity government bond is used as a measure of the risk free interest rate. The asset beta is set to 0,35 and the market premium is 4 %. On the debtside, a debt premium equal to 0,75 % is used. The WACC formula is post tax, but NVE operate with a pre-tax rate of return. The WACC-formula is rewritten to an easier formula for pre-tax calculation of the rate of return: NVE rate = 1.14Rf % (6) The formula is determined in the Regulation on economic regulation of power networks. All parameters in the WACC formula except for the risk-free rate are fixed. Any 94

97 amendments in the parameters require an amendment in the regulation. The risk-free rate on the other hand is updated annually when setting the revenue cap Minimum return The regulation states that the companies shall achieve a reasonable return on capital, given effective management, utilization and development of the network (see chapter ). It is decided that a company shall at minimum earn a return of 2 % over the last five years. Any company that falls below this minimum return will get a correction in its excess/deficit revenue balance to maintain at least a 2 % return on capital Depreciations The depreciation of capital is linear. The companies decide the economic life of components based on the conditions in their area, and then calculate the depreciation according to the expected economic lifetime of the specific network component Investments Asset additions in the last and current year (year t and t-1) are not included in the RAB as a result of the time-lag in the cost base. However, the companies can calculate the capital costs related to investments into their allowed revenue the year they are commissioned, as given in formula 3) for allowed revenue. This implies that investments are also included in the companies tariff base as of the commissioning year Network losses Network losses are measured as the difference between metered input and metered output (MWh). The costs related to the network losses are included in the cost base. To calculate the costs, NVE uses yearly reference prices of power. The prices are calculated based on volume weighted monthly area spot prices. The monthly area spot prices are collected from Nord Pool Spot AS and the weight is based on monthly national consumption volumes collected from statistics published by NVE. A mark-up of 11 NOK/MWh is added to each calculated average volume weighted area spot price. In Norway there are currently five price areas and the relevant reference price for each company is decided based on the area in which the company is operating Quality impact Introduction Incentives and penalty mechanisms are included in the regulation to assure an efficient provision of reliability of service by the network companies. A critical parameter in a credible quality regulation scheme is information about consumer valuation of different levels of reliability of supply, e.g. customers costs of interruptions. Mandatory monitoring and reporting of long interruptions (> 3 min) was introduced in 1995 and standardization of the estimation of energy not supplied in This laid the foundation for introducing quality dependent revenues and the cost of energy not supplied (CENS) arrangement in Reporting of short interruptions ( 3 min) and interrupted power became mandatory in The regulation of continuity of supply regulation in Norway was extended from 2009, and the CENS arrangement now includes all interruptions and time dependency of customers interruption costs. In addition to the CENS arrangement, NVE also introduced a direct compensation scheme for very long outages (>12 hours). The size of the compensation is decided 95

98 through the revenue cap regulation, where standard rates for compensation are specified. Any end user affected by the outage can claim compensation from the network company. However, the main aim of the scheme is to give the network company incentives to repair any outage as quickly as possible, also in cases where mainly small end users (such as households) are affected Objective for incentive based regulation on continuity of supply The Norwegian incentive based regulation on continuity of supply (CENS) gives the network companies economic motivation to ensure an optimal resource allocation when all minimum requirements are complied with. The objective is to achieve the most optimal level of continuity of supply for the society as a whole, by giving the network owners incentives to operate, maintain and develop their networks in a socio-economic optimal way and thereby provide an acceptable level of reliability. The customers costs related to interruptions are detected through nationwide surveys and will vary between different customer groups, when the interruptions occur etc. The costs related to investments to reduce the extent of interruptions will on the other hand depend significantly on the location of the customers connection to the power system, including network topology, geography, climate etc. From the regulator s point of view it is important that decisions influencing on the continuity of supply also is based on costbenefit analyses. I.e. the costs related to reduce the extent of interruptions should be lower than the future decrease in customers interruption costs due to the investment Quality dependent revenues As shown in formulas 2) and 3), the revenue caps are quality dependent, by inclusion of the element CENS. The utilities internalize the customers interruption costs. The net effect of the inclusion of the CENS element in the revenue cap and allowed revenue is that the customers indirectly are compensated for 60 % the socio-economic costs related to poor quality of supply through lower tariffs in the future. Further, the CENS element is also included in the cost norm. This implies an economic gain or loss for the specific company depending on whether the company s outage level is better or worse than the average company Method for calculation of the CENS element CENS comprises both notified and non-notified interruptions. The customers are divided in six groups: industry, commercial, large industry, public, agriculture, residential. Normalized cost data (based on a customer survey conducted in ) are used to establish continuous cost functions for each customer group. Calculation of interruption costs is based on the mandatory reporting of interruptions according to the FASIT 38 specification. The cost of an interruption of duration r at reference time, C ref, is calculated as follows: C ref = c ( r) P ref ref (7) where: 38 FASIT is a tool for collection, calculation and reporting of reliability data to the NVE 96

99 C ref = Interruption cost for an interruption at reference time c ref (r) = Cost rate in NOK/kW for interruptions of duration r P ref = Interrupted power in kw at reference time The reference time is a working day in January. The interrupted power is defined as the estimated power in kw that would have been supplied at the time of interruption if the interruption did not occur. Interruption costs are found to vary by season, weekdays and time of day. Correction factors are therefore established to adjust the calculated cost of interruption, taking this fact into account Regulation of the TSO Statnett SF is licensed as the Norwegian transmission system operator (TSO). The role and responsibilities as the system operator follows from the Regulation on system operation, where the over all goals are to facilitate an efficient electricity market with a satisfactory quality of supply. Statnett is regulated by revenue caps, as any other network company. However, there are some differences in the formula for calculation of revenue cap, mainly due to Statnetts responsibility as system operator. Furthermore, the CENS arrangement is included somewhat differently in the revenue cap for Statnett. This is because it is not possible to benchmark the CENS costs for Statnett, since there is only one TSO in Norway and TSOs in other countries do not have the similar CENS arrangement. The revenue cap and allowed revenue for Statnett are illustrated in formulas: * * RC = 0.4( C + SO ) + 0.6( C + SO ) 0.6CENS * (8) t t t t t + [( DEP DEP ) + ( RAB RAB ) WACC ] AR t = RC t + PT t + TC t 0,6CENS t + t t 2 t t 2 The cost base C t, includes the same elements as shown in formula 2), except CENS. In the revenue cap for Statnett, CENS is included as a norm (CENS*), calculated as the CPI adjusted average of Statnett s CENS costs for SO t is the actual system operation cost in year t. SO* is the system operation cost norm, calculated based on a combination of historical costs and assumptions about the cost development for future years. The SO* was fixed for a five year period as of 2008, but can be evaluated earlier if special circumstances should occur. Ct* is the cost norm, which is the result of a benchmarking analysis. The benchmarking of Statnetts transmission grid is based on the results of the international studies on the TSOs, the e3grid study 39. Statnett also owns part of the regional grid, which is benchmarked by DEA (as described in chapter 6), comparing Statnetts data to data for the remaining regional grid companies. (9) t 39 Agrell, P. J. and P. Bogetoft e3grid Final Results (2009) 97

100 The allowed revenue for Statnett includes the same elements as shown in formula 3), except the CENS element. As the CENS norm is entered into the revenue cap for Statnett weighing 60 %, actual CENS is correspondingly deducted weighing 60 % in the allowed revenue formula The Cost norm As mentioned, 40 % of the revenue cap is based on the companies actual costs as described in formula 1). The remaining 60 % is based on a cost norm. This chapter describes how the cost norm is calculated. The cost norm element has a higher weight in the revenue cap compared to the cost base to give the companies an incentive for effective management, utilization and development of the network. It is based on a benchmark analysis of the companies, where the analysis evaluates the total use of resources (input) against what the companies deliver (outputs). NVE use a Data Envelopment Analysis (DEA) method to benchmark the network companies performance. A constant return to scale (crs) approach is used. The costs and variables entering into the analysis and the cost norm are mainly based on reported data from each network company, using cost data and technical data collected through erapp, supplemented by technical data from a database for regional- and central grid components. The companies report all changes in their regional- and central grid components to NVE on an annual basis and the database is updated accordingly. In addition, NVE use data from different databases at NVE and from other external sources (e.g. the Meteorological Institute) The DEA models NVE has constructed two separate DEA models in the regulation of networks; one for the distribution grid and one for the regional and central grid (without Statnett). The input variable in both models shall reflect the total use of resources of a company, which are given by their total costs. This variable is almost identical to the cost base given in formula 2), with a few exceptions dependent on the grid level. Common for all grid levels, is that annual depreciations and return on capital related to network financed through contribution are included in the costs. The fixed assets financed through investment contribution are also included in the relevant output variables explained below. NVE has defined a set of output variables for the two models that shall describe the most relevant cost-drivers. Total costs are minimized in the analysis, given these outputs. NVE has arrived at the variables by extensive testing of different relevant variables. There are 8 cost-drivers in the model for distribution grid and 5 in the model for regional- and central grid. Some of the outputs describe the companies costs related to their structural conditions (lines and cables, transformers etc.), some are a measure of customer demand (e.g. energy delivered, no of customers) and some are constraints as geographical conditions (e.g. forest, snow). They are all cost-drivers to network companies Regional and central grid Statnett owns most of the central grid. Still some companies own small parts of the grid and the costs and output related to these parts are included in the analyses together with the regional grid. Some cost elements are excluded from the DEA analysis for the 98

101 regional and central grid. These are the costs of network losses and costs related to specific tasks for some regional grid companies. The network losses are excluded from the benchmarking because the loss volume is not only dependent of the network company s decisions, but also influenced by external factors in the network system. The output variables for regional and central grid are given in Table 1. Table 1 Output - regional and central grid Weighted value of overhead lines Weighted value of underground cables Weighted value of submarine cables Weighted value of components in stations o Switches o Transformers o Compensators Forest Distribution grid From 2010, NVE introduced a regression stage in the model for distribution grid. The DEA results from the first analysis are corrected through a regression analysis. The coefficients in the regression are a result of a panel data model, where the DEA results define the dependent variable and the variables listed in the 3 below defines the independent variables. Table 2 Output distribution grid (first stage) Energy Demand (MWh) No. Customers ex holiday cottages No. Holiday Cottages Km grid above 1kV No. Transformers Forest Snow Coastal Exposure o Wind o Coast Table 3 Variables in Stage 2 Interface to regional grid (components in station) MW Input from small hydro power stations No. islands 1 km or more from land or another supplied island One of the reasons for introducing this second stage is that a DEA model is not suitable when several variables are included. 99

102 The weight system There are 4 output variables that describe the infrastructure of the regional- and central grid companies, given in Table 1. They consist of overhead lines, underground cables, submarine cables, compensators, transformers and switches. In addition, the variable for interface to regional grid in the regression stage (Table 3) for the distribution model also includes station components that are equal to the variable for station components in the regional grid. There are several different components with different qualities and prices within each of these categories. NVE has developed a weight system that describes the relative relation between the costs connected to each component. The system refers to both capital- and O&M costs and contains 185 components of overhead lines, 44 components of underground cables, 34 components of submarine cables, 8 types of switches, 5 types of transformers, transformer capacity and 6 types of compensators. The weights have been fixed since the regulation came into force in They are based on different cost catalogues. The network components are updated annually by the companies, and registered in a database at NVE (TEK) Geographical data There are several geographical variables included in the model. Differences in geographical and climate conditions between companies may give different cost levels. For instance, a company that operates in an area with forest has to carry out forest cleaning close to their lines on a regular basis. Another company operates close to the sea and might have to maintain or replace components relatively often due to corrosion. The purpose of including geographical variables in the model is that companies with similar conditions shall be compared in the analysis. The data used in the construction of the variables are collected from different databases at NVE and from other sources of geographical data. The constructed geographical parameters are weighted with km of overhead lines, to describe how much of the network that are exposed to the specific condition. The lengths of lines are updated annually, but the geographical data has been fixed since The DEA results and calibrations Two separate analyses are carried out in each of the models. In the first analysis, the companies are benchmarked using data from two years back. The results from this analysis rank the companies efficiency from 0 to 1 using the most recent data available. The efficient companies get a DEA result equal to 1. To rank the efficient units from this analysis, we make use of historical data from the last 5 years available in a second DEA analysis. The efficient companies are here benchmarked to see if they have become more efficient, e.g. spend less resource on the same deliverables, than they did on average the last 5 years. The companies are rewarded with a DEA result larger than 1 if the analysis show improvement, if not, they get a DEA result equal to 1. As mentioned in chapter 1.2, a purpose of the regulation is to ensure that the network sector obtains a reasonable rate of return on invested capital (provided effective management, utilization and development). This purpose is obtained through a calibration of the DEA results and the cost-norm. First, the DEA results are calibrated in each of the two models, using total costs, so that the weighted average result equals 1. Then the DEA-results from both analyses are weighted together for the companies that have both distribution and regional grid, so the company ends up with an overall DEA-result. This final DEA-result is multiplied with the company s cost base to get the cost norm. The 100

103 cost norm is calibrated setting the sum of cost norms for the sector equal to the sum of the companies cost bases. These calibrations make sure that the total revenue cap for the whole sector is equal to the total cost base in formula 2) Evaluation of the economic regulation and future plans In 2010, the Ministry of Petroleum and Energy (MPE) had carried out an external, independent evaluation of the economic regulation of the Norwegian network companies. This evaluation 40 was performed by Nils-Henrik M. von der Fehr, professor at Department of Economics, University of Oslo. The main conclusions from this evaluation indicate that the economic regulation is based on approved principles, is leading in an international perspective, and is developed over time with adjustments based on experience and new challenges. The evaluation does support the economic regulation in high degree, but point at some challenges. Further, NVE has carried out several evaluations of different aspects of the economic regulation. Some have been performed by use of external consultants. Others have been performed internally, based on own expertise in combination with discussions with the electricity network industry. A major evaluation was performed during 2010, in parallel and responding to the evaluation carried out by the MPE. 41 Since the current revenue cap regulation came into force in 2007, NVE has had dialogue with the network sector regarding the companies incentives to make necessary investments. In NVEs view, the main incentives to invest should be (and are) provided through direct regulations, and not economic regulation, as described in chapter 1. However, NVE acknowledge that the grid sector find the cost norm model challenging. This challenge is mainly due to the implications of the model for cash flow from investments, as the existing model implies that a substantial part of the cash flow will be recovered late in time, as a consequence of the fact that capital costs are entered into the DEA analyses using book values and linear depreciations. This challenge was also pointed out by Professor von der Fehr. NVE is currently investigating the possibilities for accelerating the cash flow from investments, through a modification of the cost norm model. The actions under consideration do not require any major changes in the regulation. Some adjustments will therefore enter into force already from The quality of the dataset that is entered into the DEA analyses has also been subject for discussion over the past years. Challenges are connected to two aspects; both consequences of variation in the data for a single company, and consequences of possibilities of faults in the specification of the DEA model. Both Professor von der Fehr and the utilities have pointed out challenges in this respect. NVE is also considering adjustments in the cost norm model to address these issues. Some adjustments may enter into force already from 2012, others from Von der Fehr, Nils-Henrik M. Den økonomiske reguleringen av strømnettet (2010) 41 NVE dokument 13/2010: Vidareutvikling av modell for fastsetjing av kostnadsnormer for regional- og sentralnett invitasjon til innspel. 101

104 In the period after the introduction of the Energy Act in 1991, the main focus was market orientation and efficient utilization of the existing network. These elements are still important, but in addition there is an increasing attention towards the need of new- and reinvestments in the transmission system. This is regarding both an increase in the production of renewable energy and the need for reinvestments in the existing electricity network. The principles laid down in the Energy Act are not amended to facilitate such a priority, but some adjustments in recent years have been done. Among these; requirements for the network companies to connect all new production to the grid, if they are socioeconomical beneficial projects, replacement of all meters with smart meters within the end of 2016, increased economic incentives to prevent interruptions by introduction of direct payment to costumers that experience disconnections lasting more than 12 hours and by implementing short interruptions (less than 3.min.) in the economic regulation. A continuous evaluation of the economic regulation is undertaken to make sure that adjustments according to new requirements are done when it is found necessary. There is an increased focus on smart grids also in Norway, but the focus might not be as large as in other European countries, in the sense that the existing grid structure has already been built to take care of variable generation as hydropower. NVE will aim at increasing its activity regarding monitoring the network companies compliance with the new requirements and direct regulations References Agrell, P. J. and P. Bogetoft e3grid Final Results (2009) Energy Act of 1991 (LOV ) Energy Act Regulation (FOR ) NVE dokument 13/2010:Vidareutvikling av modell for fastsetjing av kostnadsnormer for regional- og sentralnett invitasjon til innspel Regulation governing financial and technical reporting, revenue caps for network operations and transmission tariffs (FOR ) Von der Fehr, Nils-Henrik M. Den økonomiske reguleringen av strømnettet (2010) 102

105 11 Country report: Sweden 11.1 The objective of the regulation Introduction Grid activity is regulated in addition to the usual rules of company law, tax rules also by special legislation on grid activity contained in the Electricity Act (1997: 857). The new ex ante regulation beginning in 2012 has focus on the revenues. The regulator shall decide on each grid companies revenue caps after a proposal from the grid companies. The revenue cap shall cover reasonable operational costs and a reasonable return on the assets used in the distribution/transmission. Quality norms are integrated in the cap so if norm values for delivery (outages) are exceeded (lowered) during the regulatory period reductions (rewards) in the following regulatory period is done in the revenue cap. The purpose is to give incentives for future improvement in quality. Energy Markets Inspectorate (EI) is responsible for monitoring the energy legislation, design rules for the regulation and decide on concessions for distribution (electricity act (1997: 857). In addition, EI follows energy markets with infrastructural services. Analyse trends and, if necessary, propose amendments to the regulations or other measures so that the function and efficiency improved. EI is also monitoring the legislation on the natural gas and district heating. A goal for EI in the supervisory work over the electricity, natural gas and district heating markets is to improve the functioning and efficiency of these markets. 42 EI is the electricity grid authority, referred to in paragraph 1 of the electricity regulation (1994: 1250). A central task is the review of the electricity grid charges (network tariffs) the customers has to pay for being connected to the grid. This task includes acting as the supervisory agency for the revenues and costs. According to the Government Bill, the regulation of electricity tariffs has the purpose that the grids activities shall be carried out efficiently (Bill p.58). The regulation shall contribute to the development of an efficient electricity market Framework for the economic regulation The Electricity Act states that the revenue cap (frame) shall finance the reasonable costs for the production of the grid services and give a reasonable return on the necessary assets (Act 2009:892). According to Section 5, 1 Electricity Act shall the revenues be decided by the regulator in advance for each regulatory period consisting of four calendar years, unless there are special grounds for another period of time ( 4). The company shall submit a proposal to EI on the request on revenues for the next regulatory period. In the proposal (application) there shall be attached the data and methodology used in determining level of revenue requested (2 )

106 The revenue cap shall cover reasonable costs of conducting network activities during the regulatory period and provide a reasonable return on capital needed to carry out the activity (regulatory asset base). The quality of companies operation of the grid shall also be taken into account when deciding the revenue cap (7 ). Such an assessment may entail an increase or reduction of what is considered to be a reasonable return on own funds. This assessment of quality will be, for the first regulatory period, done as a comparison of historic data over interruptions. Interruptions in the transmission level is taken into account to the extent that the suspension does not entail liability for network licence holder, or giving the right to interrupt compensation under Section10, 10 Electricity Act. Regarding the different tariffs that each customer pays, Electricity Act is rather sparse in the formulation on how the tariffs must be designed. The formulation in the Act states that tariffs must be objective and non-discriminatory. The rule on objectivity shall be interpreted as cost reflective: each customer shall pay his part of the grid costs. The distribution of the costs among the customers is done by means of the customer use of the capacity of the grid in the situation when demand of capacity is highest (peak-load) and by the customer s use of electricity. The rule of non-discriminatory shall be interpreted as that the tariffs must be the same for the customers within a tariff group (in terms of power and voltage level) inside the distribution area. For DSO tariffs it is not permitted with a distance element in the tariffs Regulatory period The length of regulation is set to four years. According to Electricity Act 4 a regulatory period shall be four calendar years, unless there are special grounds for another period of time. (Act 2009: 892). The first period with ex ante revenue caps will be for For the TSO level the government can decide on other lengths of the regulatory period. For the TSO level, Svenska Kraftnät (SvK) the revenue period is set to a yearly interval Differences between regulation of DSO and TSO EI is under the obligation to propose a decision on the revenue cap for SvK) to the government for the regulatory period in From 2013 EI will decide the revenue cap for Svk. A revenue cap should be established in advance of each regulatory period and each regulatory period shall be four calendar years, unless there are special reasons for a different time period. The regulatory period is one calendar year in regard to revenues from the transmission grid held by Svk. 43 This is a first difference between DSO and TSO is the regulatory period. The second difference between DSO and TSO is the estimation of the capital cost. The asset base for the TSO is completely calculated from the historical acquisition values for 43 Electricity Act, Section 5,

107 the physical assets they have in use. The historical values are indexed to account for the inflation in order to reach the cost level of The explanation for using acquisition value is the lack of norm values for the TSO assets. The third difference is that SvK also has other commitments as the overall system responsibility and liability balances 44, which causing additional cost that DSOs do not have, e.g. cost for transit, cost for counter-trade and cost of system operation, primary regulation and disturbance reserve which are treated as uncontrollable cost by the regulation. The revenue cap includes even the revenue of congestion and transit. The fourth difference is the quality indicators where transmission (and subtransmission) level reports statistics over the frequency of outages, average interruption time (AIT) and non-delivered energy (ENS). The quality norm for SvK for the year 2012 is based on the average AIT and ENS for the years The fifth difference is that statistics over controllable operating cost are based on only two years ( ) where the DSOs have statistics for four year Economic regulation Ex ante regulation with revenue caps The new regulation starting 2012 determines a reasonable cost and rate-of-return before the regulatory period starts. The regulation is based on the actual grid rather than on a theoretical virtual new grid which was the case with a preceding model (net performance assessment model). The Ex Ante regulation process starts off with the grid company submits a proposal on a revenue cap for a period of four years to the regulator (EI) under the general rule 45. This proposal is examined (scrutinized) by EI. The revenue cap shall cover reasonable cost (to conduct network activities taking into account also the quality of the supply (delivery of services). 46 EI therefore need to define the concept of reasonable cost and the quality to be pursued. Reasonable cost does not necessarily mean actual costs. Grid companies with unnecessarily high costs because of low efficiency should not be allowed to pass these costs on to customers through network tariffs For the transmission and distribution of electricity the capital costs constitutes a significant proportion of the total costs due to investments in infrastructure are very large. EI has decided to calculate capital costs based on the principle of operating capital maintenance (OCM). 47 This means that the cost of capital reflects the asset's capacity to generate services rather than its age. The basis is fixed assets replacement value and a reasonable capital cost is calculated with a real annuity method. As long as the companies are using the network grid components they are assigned a capital cost for the grid in use. This means that the capital costs essentially continues to be calculated on the basis of the same basic principles and method as EI has used since 2003 in its supervision of network tariffs 44 Electricity Act, Section 8, Section 5, 4 Electricity Act. 46 Section 5, 8 Electricity Act. 47 The other principle is financial capital maintenance (FCM). 105

108 The operating cost of a grid company consists of costs for overhead networks, maintenance, operation and administration, etc. These costs are divided in controllable and noncontrollable cost. The latter type of costs can the company pass through completely. For the controllable cost an annual efficiency requirement i set to one % (1 %), which in real terms will force the company to reduce the controllable operating cost if they will keep the rate-of-return at the same level. The allowed revenue is adjusted depending on the outcome of quality in the regulatory period. The operative and capital cost are then added together to get a total revenue cap. In the Figure 1 below shows schematically how the revenue cap is determined. Figure 1 Components in the economic regulation Revenue frame Important parameters are valuation of the physical grid, a norm for the cost of capital, the depreciation times for the physical components, the initial level for operating cost and the efficiency target. Norms on quality are also to be set, measured and reviewed. After each period the outcome for interruptions is compared to the norm level set by the regulator Transition from ex post to ex ante The starting point for implementing the new regulatory model with ex ante calculation of revenue caps is based on operating costs and rate of return for reasonable effective company should have over time to manage their obligations as electricity grid company. With this revenue the company will have cost recovery and return on capital, so that the company can operate the grid with a high supply security and carry out the necessary investments for the development of electricity networks. In order to ensure that all companies are treated equally EI s method is disconnected from the companies historical accounting costs. EI s method is modelled on the principle that the company should give cost recovery over fixed assets life. Revenue cap according to the developed method would mean that the allowed revenue cap will increase about 35 % over six years compared to current revenue levels. Such increase is completely against the Electricity Act's principle of customer interest in low and stable network charges. 106

109 After an analysis of the differences between results of the method and the current revenue level, EI assessed that an appropriate time to equalize revenue caps calculated by the method used and the current revenue level are four regulatory periods (16 year). Since EI's calculations are based on the companies' revenues for the years in the price level of 2010, the first regulatory period also include the years 2010 and 2011 and therefore the number of years are 18. For the period , a company's revenue cap will be adjusted. The difference between the result of the standard method and the mean historic revenue levels and ( ) is divided by one third (or 6/18) Capital cost The capital cost is a product of a decided rate of return on assets, the calculated capital base and chosen depreciation times. The main principle to calculate the regulatory asset base is to use replacement values with real annuities. 48 This valuation principle is supplemented by some other methods when explicit values of replacement are not known Asset base The basis for calculating capital cost is fixed assets replacement value. By fixed assets means the physical equipment that the grid company uses to operate the grid. Other real capital equipment like office buildings, vehicles, computers which the company uses but are not directly designed for distribution of electricity are converted to an operating cost. To determine the regulatory asset base there are four methods which may be used to value assets within electricity networks. EI's report EI R2010: 07 on valuation of electricity grid real assets set four methods (see also EI's PM 2010: 11 and 12, 2010: 2010: 13). For more information about the accounting of capital base is set out in Section 5 in EIFS2010: 6 (Regulation on the transmission of data). The four methods for valuation listed below shall be used in descending order. That is standard values is the default alternative, then aquisation values, then book-keeping values and last some other valuation. To adjust for inflation during the regulatory period the revaluation of assets will be done by an index from statistics Sweden (SCB). The index in question is an index for the development of production cost for buildings. Valuation according to standard values Fixed asset included in the capital base shall as a first alternative be set to a replacement value that corresponds to a standard value. A standard value on each equipment to be calculated on the basis of the investment expenditure a grid company would have to acquire or produce in a cost-effective manner with due regard to conditions which grid company cannot affect. EI has decided on approximately 145 standard values for facilities 48 Förordning (2010:304) om fastställande av intäktsram enligt ellagen (1997:857). 49 Energimarknadsinspektionen, Värdering av elnätsföretagens kapitalbas i förhandsregleringen, EI R2010:07, juni

110 with voltage up to and including 24 kv and about 500 standard values for facilities above 24 kv up to 220 kv. 50 For instance there have been decided different norm values for investment in cables depending on the character on the soil and urbanisation. 51 Valuation according to value at investment time If the standard value is missing for a facility, the rules states that the value may be calculated on the basis of the expenditure for the acquisition or manufacturing of the fixed asset when it originally came into operation in the network activities taking into account the change in price mode from the acquisition. The company must be able to verify that the cost is the original. Valuation according to book value If there are no prerequisites to compute a replacement based on the standard value or acquisition value, cost is calculated on basis of the asset's book value. Valuation by other way If none of the above methods are not possible to use, the value of the facility has to be determined by a reasonable method 52. An example could be a rented facility that is not in the list of standard values Return-on-capital (cost of capital) Weighted Average Cost of Capital (WACC) is an established method for determining the rates of return in a business. WACC is used by several regulators for deciding on the norm value of rate of return. The aim for deciding on a rate-of-return is to estimate the level that is necessary for attracting capital to the industry. If the rate-of-return is set higher than the necessary level, there can be too much investments in the industry relative to other industries and the customers has to pay more than necessary for their services. A reasonable rate-of-return shall support an effective investment level and thereby give customers low tariffs both in short and long run. For the first regulatory period WACC will be used as method for deciding the norm value for the rate of return. EI has asked two financial consultancies on their expertise to give an interval (minimum to maximum) for a reasonable interest. The discussion has been focussed on how and how much the interest shall be influenced by the fact that the companies have untaxed reserves. EI commissioned an estimation of a WACC for the regulatory period. The two reports were analysed and discussed and as a result there were some changes in the parameters. In Table 1 an overview of the parameters for the WACC. The finally decided WACC was set to 5,2 % for the first regulatory period without any change during the period Three classes of environment: city, rural and rural with complicated soil conditions of the decree. 53 Energimarknadsinspektionen, Kalkylränta i elnätsverksamheten, PM 2011:07, september

111 Table 1 Rate of return parameters for the first regulatory period Risk free rate nominal 4,0% Market risk premium 4,74% Tax rate 20,0% Asset beta 0,38 Equity beta 0,66 Illiquidity premium 0,50% Debt share 0,50 Debt/equity share 1,00 Expectation on inflation 1,99% Debt premium 1,49% Operating cost Operating cost is the other part of the total cost for the grid services. The operative cost can in turn be divided in two parts: controllable and non-controllable. The latter part is cost which the company has no or little control over in terms of possibility to reduce over time. The first part is costs which the company has the possibility to reduce for given services (output). In a special report EI analysed how the division between different cost items should be done The level of operating cost The starting point is the reports from the DSOs their historical costs taken from audited data from bookkeeping. In theory should the latest year with actual historic data be the starting value for assessing the revenue frame. But to avoid that this year can be an extreme year (with high or low cost for the services), the starting value will be set to average of the four years and inflated them to the price level of Non-controllable operating costs Cost that is defined as non-controllable is possible for the company to pass through to the customers. Those costs do not have any efficiency requirement. The most typical kind of such cost is the charges the DSO has to pay for to be connected to subtransmission level. The grid losses are controllable in the long run. In the short run (the regulatory period) are the physical losses non-controllable. But the tendering of electricity for these losses can the DSOs have influence over in the short run. The physical losses in terms of electricity are not possible to control. In the medium and long term the losses are possible to control both by the tendering and by investment in the grid. The grid losses will for the first period be seen as non-controllable cost. The Electricity Act states that the electricity for covering grid losses shall be tendered in a competitive way. The grid company has to 54 Energimarknadsinspektionen, Löpande kostnader i förhandsregleringen, EI R2010:06, april

112 tender the electricity for the losses in an open, non-discriminatory and market orientated way. 55 This cost is possible to examine in a revision outside the economic regulation of revenue caps. So, cost for losses is of cost-pass-through type but the tendering of electricity is possible to examine and control that the DSO buy their electricity on market based conditions. This is decided for the first regulatory period. The payment for feeding in electricity from generation plants is also considered as an uncontrollable cost. This payment to the producer shall be a compensation for the benefit the grid owner has for this feeding in of electricity (less losses of energy on the grid and less payment to upstream grid). These costs shall be accounted in a separate way and separated from the controllable cost. The grid company shall make a forecast on the expected cost for the regulatory period. After the end of the regulatory period these cost shall be compared and adjusted if necessary Controllable operating costs The concept of controllable costs is necessary because those costs will have an efficiency target. The degree of control can vary between companies depending on what type of cost and contracting of the resources. The common denominator is that there are possibilities to reduce these costs over time. Especially in connection with mergers potentials for reducing costs appear. EI has analysed and defined the cost items in controllable and noncontrollable. 56 A typical controllable cost is the cost for the staff and services bought. Some capital cost as for owned building (not defined as grid assets) or vehicles are converted to a controllable cost. Assets that are leased by the company are regarded as a part of the regulatory asset base and converted to a capital cost The efficiency target EI has decided that a general requirement shall be applied on current controllable costs. To determine the size of the efficiency requirement, several analyses have been made. Productivity developments have been studied by calculations for the Swedish electricity grid companies, through comparisons with other sectors and by international comparisons. 57 Development of productivity in the years 2000 to 2008 shows a trend of increases. The costs of production in terms of making capacity available for electricity customers have decreased by approximately 2% per year. The estimated productivity developments do not from what several other studies indicate. The various studies of the development made gives different results depending on the choice of method, model, and year. Development of productivity is affected for example of the form of regulation. The deregulation of the electricity industry for the last 20 years has often been a change from a return-on-capital-regulation to different kinds of incentive regulation. The reason for this change was to provide incentives for the grid companies to be more cost efficient (productive). This means that studies which relate to a period of rate-of-return regulation (cost-of-services regulation) should show a relatively slower development compared with 55 Ellagen 1997: Energimarknadsinspektionen, Löpande kostnader i förhandsregleringen grundprinciper vid beräkningen, EI R2010:06, april Energimarknadsinspektionen, Förhandsregleringen krav på effektiviseringar intäktsramen för löpande kostnader, EI R2010:

113 the period of incentive regulation. In connection with the restructuring of the electricity market and the introduction of incentive regulation has often great structural changes occurred. These structural changes will drive on productivity developments. This means that studies involving a period before and after a reregulation should show high productivity development. To sum up, the different studies show improvements in productivity over time. On the basis of the studies on Swedish data found that productivity developments for the years has been in the range of 1,5% to 3,5% per year, depending on the choice of method, model, and data. Other studies show the average annual increases of around 2,7 % as average. Assuming a continued gradual increase in productivity, the cost per unit of service will decrease over the years. In recent years, partly a technical change occurred with the shift from lines to cables. This is expected to provide both decreased maintenance costs and fewer interruptions. The decision for the efficiency target for the first period founded on empirical studies and reviews of other productivity studies was to set the general requirement on operating controllable cost to 1% per year Quality regulation Security of supply regulation is important in incentive regulation because there can be an unintended negative impact on quality of supply when the DSO must be more efficient to keep the rate-of-return on the same level. Norms on quality is therefore essential for the customers. Norms can be set to force the DSO to produce a better service. Because grid companies regulatory asset base is calculated from the existing physical assets valued today at the reinvestment value and the costs of operation and maintenance are based on historical costs, there will be incentives to reducing maintenance in order to increase profits. Without a specific regulation of the supply security, the companies can increase their profits in the short term by reduce reinvestment and running costs for the operation and maintenance. This will eventually in the long-term result in a deterioration of security of supply as a result. A quality component of the regulation is therefore necessary to provide incentives to enterprises to networks at least maintain, but also improve the quality of operations. It is not possible to apply a system of interruption statistics per customer for the first regulatory period because such statistics are missing at the moment. Through a brief analysis of the available customer interruption assessments (cost for the customers) as well as by studying the application of quality control in the rest of Europe, the decision by EI is to use the cost assessment study done by the Swedish Energy (the industry association for DSOs). This study was carried out in Swedish Energy did a simple update of these interruption costs for customers in 2003 by inflating the values by change in consumer price indices (CPIs). 58 The calculation relating to the outcome of 2012 will therefore be adjusted to the price level in 2012, and so on. 58 Swedish Energy is the industry association for the DSO in Sweden. 111

114 In Section 5, 7 of the Electricity Act states that quality deduction in the revenue frames shall be limited to a maximum of an amount corresponding to the return on the capital base. The rule has been added to protect mainly smaller network companies in case of extreme weather impact. EI will therefore limit the annual amount of quality adjustment for the period ( ), by introducing a ceiling and a floor as maximum amounts to return on own funds, but not more than 3 % of annual revenue. The introduction of the ceiling on how high a supplement can be detrimental unjustifiably high supply security while a floor would restrict the negative economic consequences for network operators The measurement of quality A central starting point of the new regulation is to define a relationship between the economic benefits and network companies ' performance in respect of the supply security. In practice, this means that the DSOs are rewarded if they meet the intended objective of good supply reliability, but a financial deduction if the reliability of supply will be lower than the standard level. Figure 2 illustrates how this connection could be drawn up. Since 1998, Inspectorate collected interrupt statistics from all local network companies for interruptions longer than 3 minutes. The interrupt indicators used is SAIDI (System Average Interruption Duration Index) and SAIFI (System Average Interruption Frequency Index). Interruption statistics, SAIDI and SAIFI is divided into announced outages (planned) and unplanned interruptions. It s not possible today to distinguish interruption statistics in various intervals, from 3 minutes to 12 hours and from 12 hours to 24 hours from the interrupt data collected. The interruption statistics that has been gathered the TSO and subtransmission level (regional grid operators) are measured in terms of interrupt frequency, average downtime (AIT) and the Non-delivered energy (ILE). The interrupts are divided into three intervals, kv, 130 to 70 kv, and greater than 130 kv The impact of quality on the revenue frames Unplanned outages between 3 minutes and 12 hours and planned outages longer than 3 minutes will be used as a basis for quality control. The reason for this restriction is that network operators are obliged to pay compensation for outages that are longer than 12 hours, which means that there is already an incentive to reduce these interruptions. For DSO, the indicators of supply security for the first supervisory period will be medium interruption duration, SAIDI (System Average Interruption Duration Index), and medium frequency interruption, SAIFI (System Average Interruption Frequency Index). These indicators are established and there are statistics for several years. For regional networks the appropriate quality indicators are ratios over annual average non delivered energy (ILE) and annual average non delivered power (ILEffekt) calculated per withdrawal point for both planned and unplanned interruptions. These indicators are 59 Energimarknadsinspektionen, Kvalitetsbedömning av elnät vid förhandsregleringen, EI R2010:08, maj

115 established and used by the region's companies. Indicators are also commonly used in quality control in other European countries. Grid operators have different objective environments for maintaining supply security. It is therefore not possible to require the same level of supply security. The regulation has to take care of these differing environments. In the regulation for the first period each DSO is compared by itself because the norm of quality is founded on the historically statistics for each DSO. Norm levels of security of supply are determined for each individual network company for the first period. These levels are calculated as average of the four years If the DSO can keep this standard level during the first regulatory period, no adjustment of the revenue framework will be done in the follow-up of the outcome of the regulation. Figure 2 illustrates the principle for the quality regulation in the assessment of the allowed revenue. Figure 2 Quality impact on allowed revenue Increase Maximum increase Low quality High quality Maximum decrease Reduction The maximum increase of better quality than the initial norm values is set to 3 % of the revenue cap and the other way around if the outcome of quality has been worse. The slope of the line of 45 grades means that half of the better quality is given to the customers in terms of better quality and the other half goes to the company as a reward for better quality. That is, half of the increase in quality relative the starting value will give the company a supplement in the revenue cap for the next regulatory period The quality in the future regulation For the second period, it s possible to develop a better model for quality control. The most important component of this is the high level of detail in the new interrupt reporting. There are also opportunities to carry out a new survey of consumers' valuation of interruption (willingness to pay/willingness to accept studies). Conditions to calculate a model that governs the DSO to a social optimal quality level should then be possible. 113

116 The regulator, EI will therefore initiate work on developing quality adjustment in the second period rules during the spring A development project should begin on the quality of the interrupt statistics, categorisation of customers and customer business continuity assessment and other factors that may be relevant for the assessment of the quality of voltage quality and quality of customer service Applied benchmarking methods The regulatory model in use has a general efficiency target for productivity development. So there has not been any benchmarking of the DSO in order to get a basis for setting individual efficiency targets. For the first regulatory period with ex ante there was a decision by the regulator to only use a general X-factor or efficiency target. The revenue cap for controllable operating cost in real terms will be 1 % lower each year for the period. The grid company who can reduce the operating cost with more than 1 % per year have possibility to lower tariffs and/or get a higher return-on-capital. The efficiency target applies to both DSO and TSO level. An analysis of the productivity development for the years in the DSO sector in Sweden was done with application of both regression analysis, SFA and DEA. The mean development of productivity was estimated to 2 % per year for this period. Other studies of the productivity development were also used as information. 60 The model used for the estimation of productivity development consisted of controllable operative cost as input and three outputs (number of customers, length of lines and cables and installed capacity of transformers. There was no need to incorporate contextual variables because the focus was to estimate the development of productivity and not to make comparisons between the companies. Earlier models for bench marking used to incorporate the volume of electricity distributed, which is common in other models. The reason for not including distributed electricity is that the grid losses in not included in the operating cost Issues in future regulation The methods now developed and applied for the period will be further developed. Of course the experience and coming evaluations of the outcome for the first regulatory period will be of essential value for future changes. But there are already some aspects which will be analysed. The quality regulation must have more precision (sharper) which will be possible with better statistics on outages. For the second regulatory period, the prerequisites to apply a much more developed quality regulation are present. Then the input for the determination of reference level will have been improved radically. The following factors will have much more reliable data: 60 Energimarknadsinspektionen, Förhandsregleringens krav på effektiviseringar intäktsramen för löpande kostnader, EI R2010:11, juni

117 data on outages for each load point data on yearly mean of delivered energy (ILE) and yearly mean not-delivered power for each load point customer interruption cost rating for different customer groups With detailed outage data for a five-year term ( ) there are very good conditions to determine the reference levels for a number of different categories of customer. A new questionnaire will be done where customers valuation of outages are estimated. The value of lost load can be estimated by means of methods of willingness to pay (or accept). This opens for differentiated interruption values for different customer groups. EI therefore intends in the spring of 2012 begin work to develop this issues. Another aspect which will be addressed is the development of norm cost on operating cost. There was a discussion before the first regulatory period to use an existent list of norm cost for operating cost coupled to the different equipment in the capital base. The analysis of these norm cost showed that the existing list of norm cost is not reliable. The researchers recommended a thorough work to develop these norm costs. 61 EI decided not to use these norm cost in the in the first period. A third aspect to be studied before the second regulatory period is the revision of the efficiency target. How has the productivity developed during the first period? How shall the future efficiency target be set? Will it be a general target, as a firm-specific target, or as a combination? What methods should be used? Experience from other regulator A fourth aspect to develop is the list of standard values for the grid equipment References Förordning (2010:304) om fastställande av intäktsram enligt ellagen (1997:857). Energimarknadsinspektionen, Värdering av elnätsföretagens kapitalbas i förhandsregleringen, EI R2010:07, juni Energimarknadsinspektionen, Löpande kostnader i förhandsregleringen grundprinciper vid beräkningen, EI R2010:06, april Energimarknadsinspektionen, Kvalitetsbedömning av elnät vid förhandsregleringen, EI R2010:08, maj Energimarknadsinspektionen, Förhandsregleringens krav på effektiviseringar intäktsramen för löpande kostnader, EI R2010:11, juni Energimarknadsinspektionen, Kalkylränta i elnätsverksamheten, PM 2011:07, september Hilber P, Isenberg A, Setreus J and Wallnerstörm C-J, Potentiell användning av standardkostnader i regleringen av elnätsföretagens löpande påverkbara kostnader, KTH TRITA-EE 2010:059, dec Hilber P, Isenberg A, Setreus J and Wallnerstörm C-J, Potentiell användning av standardkostnader i regleringen av elnätsföretagens löpande påverkbara kostnader, KTH TRITA-EE 2010:059, dec

118 116

119 NordREG c/o Danish Energy Regulatory Authority Nyropsgade 30 DK-1780 Copenhagen V Denmark Telephone: (+45) Telefax: (+45) [email protected] Internet: