Interconnector participation in Capacity Remuneration Mechanisms

Interconnector participation in Capacity Remuneration Mechanisms A REPORT PREPARED FOR ENERGY NORWAY January 2014 Frontier Economics Ltd, London.

January 2014 Frontier Economics i Interconnector participation in Capacity Remuneration Mechanisms Executive Summary 3 1 Introduction 7 2 The current context 9 2.1 The development of a CRM in the GB market... 9 2.2 The development of a CRM in the German market... 16 3 Impact of CRMs on market pricing 18 3.1 Impact of CRMs on pricing... 19 3.2 Impact of CRMs on interconnectors... 21 4 Arguments for the participation of interconnected capacity in CRMs 24 4.1 Security of supply... 24 4.2 Economic Efficiency... 25 4.3 Scale of economic efficiency impact... 29 4.4 Potential European Commission viewpoint... 32 4.5 Conclusion on including interconnectors in CRMs... 33 5 Developing options for inclusion 34 5.1 Securing the flow... 34 5.2 Enabling participation: strawmen... 37 5.3 Assessing the options... 44 6 Conclusions 51 Contents

ii Frontier Economics January 2014 Interconnector participation in Capacity Remuneration Mechanisms Figure 1. The recovery of fixed investment costs through mark-ups to SRMC 11 Figure 2. Stages of proposed CRM in UK market 13 Figure 3. Capacity market has the potential to reduce wholesale market revenues 19 Figure 4. Observed prices without a CRM in the GB, German and Norwegian markets 22 Figure 5. Simulation of new link between GB and Norway in 2025 23 Figure 6. No distortion to trade over the interconnector 26 Figure 7. Distortion to trade over the interconnector leading to a welfare loss 27 Figure 8. Welfare effect from short term gains in trade due to IC expansion 30 Figure 9. Matrix of options one to four 39 Table 1. Eligible and ineligible capacity in the CRM auction 14 Table 2. CRM options in operation and under consideration in Europe 18 Table 3. Scenarios for price distortions due to CRMs 44 Table 4. Summary of options assessment 48 Table 5. Summary of questions and answers in relation to top ranked option Auction 49 Table 6. Summary of questions and answers in relation to top ranked option Delivery 50 Tables & Figures

January 2014 Frontier Economics 3 Executive Summary There is a movement in a number of European markets towards the implementation of generation Capacity Remuneration Mechanisms (CRMs). The debate is at different stages according to Member State, with CRMs implemented in some, such as Ireland and Spain, and well developed plans in others such as Great Britain (GB). This shift is driven by a concern that the incentive to invest in conventional thermal generation is being undermined by the (politically driven) rapid growth in renewable powered generation. Renewables are lowering the wholesale price, and reducing the number of hours over which thermal mid-merit and peaking plant can make a return. As a result thermal plants are reliant on very high prices during periods of scarcity, increasing revenue uncertainty and raising the risk of regulatory intervention in response to a perceived abuse of market power. These factors, combined with uncertainty as to the future pace of renewables build out, are increasing the risk premium for investment or deterring new investment entirely in the thermal plant required to support the power system when intermittent renewable generation is not running. The movement towards capacity markets in GB and Germany could impact on Norwegian interconnectors planned to connect with those countries. Whilst Germany is as yet undecided about whether to implement a CRM, GB plans to implement a centralised capacity market in 2014, with the first delivery year in 2018/19. As currently proposed, interconnected capacity is not able to bid into the CRM auction. This report argues that it is beneficial to include interconnected capacity in a CRM and sets out a possible way forwards for doing so. Arguments for including interconnected capacity in CRMs In assessing whether interconnected capacity should be included in CRMs we consider arguments relating to security of supply and economic efficiency (operational and investment). Security of supply interconnectors are physically as reliable as local generation, however there are potential concerns about the reliability of the direction of flow for an interconnector during a period of power system stress (i.e. a period with a very tight demand supply balance). This might be due to a coincident stress event in the neighbouring market or due to market imperfections, in particular in the cross-border intra-day or balancing markets. These factors need to be addressed in the design of options for including interconnected capacity in CRMs. Executive Summary

4 Frontier Economics January 2014 Operational economic efficiency in terms of operational efficiency there are good reasons to expect wholesale energy prices in the local market to be lower due to the presence of a CRM. For example, increased competition at each demand level, and an increased threat of regulatory intervention could lead to pricing closer to short-run marginal cost (SRMC). Generators outside of the CRM in the neighbouring market may price on the basis of SRMC plus a mark-up, in order to recover plant fixed costs. If this is the case then there is the potential for a welfare loss due to more expensive generation in the local market being selected to generate in preference to lower SRMC generation in a neighbouring market. However, despite this welfare loss, including interconnected capacity in the CRM is unlikely to lead to a welfare gain in terms of short term operational efficiency. Some or all of the value from the capacity payments that accrue to the neighbouring market is likely to pass through to the interconnector owner. As a result the bidding behaviour of interconnected generators is unlikely to change from SRMC plus a mark-up. Investment efficiency the key efficiency argument for including interconnected capacity in CRMs rests with investment efficiency. Excluding interconnectors or non-domestic generators from participating in a CRM would skew investment signals in favour of local generation. Even if it were cheaper to provide capacity to meet security of supply standards by developing more interconnection capacity, investors may choose to develop new local power plants due to the additional investment incentive provided by the CRM over and above that provided by the energy price. The effect would be to raise the cost of meeting the security standard if the lowest cost sources of capacity (across the set of possible generator and interconnector projects) were not developed. Conversely, including interconnectors or nondomestic generators in the CRM would allow the interconnector owner to capture the reliability value of the interconnector, giving them the equivalent investment signal as is given to local generators. There would appear to be a strong case for including interconnected capacity in CRMs on the grounds of economic efficiency. The EC s current thinking with respect to CRM design in the context of State Aid legislation appears to also favour including interconnected capacity in CRMs. Developing options for inclusion We develop and assess five options for including interconnected capacity in CRMs. In doing so, we assume that two key design principles apply to each of the options. These two principles represent a way forward to tackle the important security of supply concern about the reliability of interconnectors at times of system stress: Executive Summary

January 2014 Frontier Economics 5 The interconnected capacity that can bid into the CRM is de-rated to reflect not only physical reliability but also the risk of coincident stress events. This should be done on a probabilistic basis reflecting the historic occurrences of coincident stress events in the neighbouring markets. The Electricity Target Model will be implemented to ensure that interconnected capacity can respond efficiently to price signals across borders. Policymakers have an important role to play in ensuring that coupled intra-day markets are allowed to develop providing a trading route for generators to respond to a stress event in the CRM market. Greater inter-tso cooperation in relation to balancing will provide a second opportunity for the market to respond. The implementation of arrangements consistent with these two principles should go a long way to addressing concerns as to the risk that power does not flow to the right market in times of system stress. Their implementation is largely independent of the approach taken to including interconnected capacity in the CRM. We set out and assess five different options or strawmen for including interconnected capacity in CRMs. The first four options are closely related and differ with respect to the party bidding into the CRM and the allocation of risk of non-delivery of power by the interconnector. The fifth option is an alternative to the participation of interconnected capacity in the CRM, and instead focuses on removing the potential distortion created by the CRM to energy market prices. The key points from our assessment of the five options are: Some or all of the value from the capacity payments should flow either directly or indirectly to the interconnector owner depending on the option chosen. If the interconnector owner bids directly in the CRM auction then they will receive the full value of the capacity payments. Alternatively if the generators bid, much of the value could flow to the interconnector owner through a gateway auction, where successful bidders receive a ticket allowing them to bid into the CRM, the price of which would represent the future capacity payments bidders expect to receive. The auction for this ticket could be linked to existing auctions for physical or financial transmission rights or be separate. The quantity of rights in the auction is determined by the de-rated capacity of the interconnector. There are good arguments for generators being the parties bidding into the CRM, rather than the interconnector owner. This would still support investment in interconnectors, but in addition there is a strong incentive on those generators located abroad with CRM contracts to respond to price signals. This could be either through responding to bids in the intra-day Executive Summary

6 Frontier Economics January 2014 markets or bidding into the balancing mechanism. This would increase the probability that the interconnector flows power to the CRM market during periods of system stress. Investment efficiency is stronger in the options where the full risk of nondelivery at times of stress is placed on the interconnected capacity. This should reduce costs to consumers in the CRM market. Under these options, the interconnector owners or the generators are exposed not only to the risk of their physical reliability, but also to the risk of coincident stress events (even if this is partially outside of their control). By internalising this risk, investment decisions should correctly reflect the true reliability value of the interconnector, ensuring the correct balance between local generation and interconnector investment. The fifth option assessed involves an additional payment to all imports reflective of the capacity value of the interconnector. Whilst in theory this could correct a market failure, it provides a weaker interconnector investment signal than receiving a stream of fixed capacity payments directly, possibly leading to less investment. This option attempts to correct for a distortion in the market price, so in theory dispatch efficiency could be improved. However this is unlikely, with the most likely outcome being that new distortions to operations and dispatch are created. Executive Summary

January 2014 Frontier Economics 7 1 Introduction There is a movement in a number of European markets towards the implementation of generation Capacity Remuneration Mechanisms (CRMs). The debate is at different stages according to Member State (MS), with CRMs implemented in some, such as Ireland and Spain, and well developed plans in others such as Great Britain (GB). There are a range of different approaches being adopted or considered, such as quantity obligations, capacity payments and strategic reserves, leading to concerns that the interventions are too nationally focused and do not consider cross-border trade in their design. The implementation of CRMs has the potential to affect the development of interconnectors, and with it a more integrated single European market. CRMs may affect interconnector investment in the following ways: impact energy prices, and hence the value of the interconnector capacity; provide a potential new source of payments to the interconnector owner or users; and, create a differential between the treatment of interconnector capacity and national generation capacity, reducing the value of interconnector capacity relative to that of local generation. Energy Norway, a non-profit industry organization representing about 270 companies involved in the production, distribution and trading of electricity in Norway, commissioned Frontier Economics to consider the practical implications of the development of CRMs for Nordic interconnectors. The most advanced projects in Norway are for interconnectors with GB and Germany, so there is a particular emphasis on the development of CRMs in these markets. Plans in Germany are at an early stage, so the focus of the discussion is largely based on a centralised capacity market which is being developed in GB. We have structured the report as follows: in section 2, we describe the European context in more detail and in particular summarise the rationale for and the design of the proposed CRM in GB; in section 3, we set out the theory behind the impact of CRMs on market prices and hence potential knock-on impact for interconnector investment; in section 4, we set out the issues for the inclusion of interconnectors in CRMs, with a focus on security of supply and economic efficiency; Introduction

8 Frontier Economics January 2014 having concluded that there are material welfare benefits to be gained by their inclusion, in section 5, we describe practical options for the integration of interconnectors into a CRM, and based on a set of assessment criteria, the option that would appear to be the most appropriate model for development. Introduction

January 2014 Frontier Economics 9 2 The current context Currently in Europe there are a large number of Member States that are considering intervening in their electricity market to ensure security of supply. Currently seven countries have introduced a CRM and many others, including Germany and Great Britain (GB), are considering doing so. This shift is driven by a concern that the incentive to invest in conventional thermal generation is being undermined by the (politically driven) rapid growth in renewable powered generation. Renewables are lowering the wholesale price, and reducing the number of hours over which thermal mid-merit and peaking plant can make a return. As a result thermal plants are reliant on very high prices during periods of scarcity, increasing revenue uncertainty and raising the risk of regulatory intervention in response to a perceived abuse of market power. These factors, combined with uncertainty as to the future pace of renewables build out, are increasing the risk premium for investment or deterring new investment entirely in the thermal plant required to support the power system when intermittent renewable generation is not running. A recent paper by ACER 1 highlighted a growing concern that countries are pursuing CRM policies in a nationally focussed way, choosing diverging approaches, and are typically devoting limited attention to the impact of these mechanisms on cross-border trade. Given the highly interconnected nature of many European markets, and a future vision of a single Internal Electricity Market for Europe this is a key concern. Potential impacts include: short-run prices can be affected, distorting least cost dispatch; long-run investment decisions could be shifted away from more costeffective locations; and the cost of a CRM is borne by the consumers of one Member State but there are spillover benefits to neighbouring markets. This section goes onto look at the GB and German markets where a CRM is being considered in the case of Germany and implemented in the case of GB. 2.1 The development of a CRM in the GB market The UK Government as part of its Electricity Market Reform proposals has decided to introduce a CRM with the first auction planned for the end of 2014. The capacity procured via the auction will be obliged to deliver in time for winter 2018/19. There is still some uncertainty over the final proposals as they are 1 Capacity Remuneration Mechanisms and the Internal Market for Electricity, July 2013, ACER The current context

10 Frontier Economics January 2014 subject to UK parliamentary and EU State Aid approval. However there is considerable detail available on the proposed design which is set out in the remainder of this section. A key feature of the current design is that non-gb capacity will not be able to participate in the first auction in 2014. The Department for Energy and Climate Change (DECC) states that if a suitable model can be developed for its incorporation then non-gb capacity will be able to bid into future auctions. DECC recognises the potential efficiency gains from increasing competition in the auction as well as providing appropriate incentives to invest in interconnection. However, as yet they have not been able to produce a model that preserves the integrity of the capacity market itself and is compatible with European internal energy market rules. 2.1.1 Rationale for intervention DECC identified a number of concerns with the current model of an energy only market that have led to the conclusion that a CRM was necessary to maintain security of supply. These arguments are not unique to the GB market and form the basis of intervention in many European markets. Energy-only market and security of supply In an energy only market, peak load pricing will in theory lead to an optimal level of investment in new capacity. For the majority of time periods in a competitive wholesale electricity market there will be more generating capacity available than required to meet demand. So the market price will typically reflect the short-run marginal cost (SRMC) of the most expensive generator required to balance the system. As a result, all generators with lower operating costs receive an inframarginal rent largely covering their fixed costs and helping to support investment in new baseload and mid-merit capacity. However, this may not provide sufficient remuneration to support investment in baseload and mid-merit capacity and nor will it ensure that at times of peak demand the thermal peaking plant which typically have the highest short-run costs of generation receive enough remuneration to support their investment. Therefore, at times of peak load when the margin between the available capacity and demand tightens, prices will exceed the marginal costs of the most expensive plant on the system, leading to a scarcity rent or mark-up. Even at demand levels below the supply capacity, as demand tightens, competition will be lower and so mark-ups will develop. This mark-up provides an essential signal to investors in peaking plant, helps to support investment in baseload and mid-merit plant and provides a strong incentive for generators to be reliable so that they capture the peak prices. This is illustrated in Figure 1. The current context

Price, Costs January 2014 Frontier Economics 11 Figure 1. The recovery of fixed investment costs through mark-ups to SRMC Demand + net exports SRMC + markup P* SRMC MWh Inframarginal rent Potential Imperfections in the energy only market DECC argues that there are a number of potential reasons as to why this mechanism may break-down leading to underinvestment in peaking capacity in the electricity market: Uncertainty and volatility a potential investor is reliant on earning a return from high prices over a relatively small number of hours. The actual level of prices that prevail and the number of hours which yield a scarcity rent is highly uncertain and difficult to forecast. This uncertainty will increase investment risk premiums increasing overall costs, or deter investment all together. Political risk the uncertainty is likely to increase as the market develops to include a greater share of intermittent renewable technologies with a very low short run marginal cost of production. Remuneration for thermal plant and in particular for flexible peaking capacity will become increasingly reliant on very high prices in even fewer hours. This risk is heightened further given the uncertainty about the pace of investment in renewables and low carbon technologies. Missing money there are a number of reasons why the electricity market may not sufficiently value capacity at peak times: The current context

12 Frontier Economics January 2014 GB imbalance prices do not currently sufficiently reflect the value of capacity at times of scarcity. This is because the charges placed on generators for being out of balance do not sufficiently reflect the full cost of the balancing actions taken by the System Operator. Regulatory risk that prices won t be allowed to rise. Investors in thermal plant are concerned that the regulator/government will intervene in the energy market on the basis that very high market prices represent an act of market power e.g. by introducing a price cap. As a consequence of these issues, the UK government has taken a view that the market will not provide sufficient signals for investment in additional thermal capacity, or provide sufficient incentives for plant to be reliable. DECC has proposed that the first delivery year for the CRM should be 2018/19, which follows from the earliest possible implementation date of the mechanism. Predicting exactly when capacity problems may arise in the market is extremely difficult due to the inherent uncertainties in forecasting future demand, future investment and retirement decisions and the future direction of interconnector flows at times of stress. It is therefore possible that no additional capacity is required in the first auction, meaning the price will be low. However by introducing the capacity market as soon as possible the intention is to mitigate the potential risk of insufficient capacity and reduce regulatory uncertainty in the market. 2.1.2 Key features of the UK CRM design The following section summarises some of the key design features set out by DECC, in particular highlighting the features of most relevance to the subsequent discussion regarding the inclusion of interconnected capacity into the CRM. DECC has set out their proposed design in five operational stages which provide a useful framework for describing the mechanism, as shown in Figure 2. The current context

January 2014 Frontier Economics 13 Figure 2. Stages of proposed CRM in UK market Capacity to procure Auction Trading Delivery Payment National Grid provides advice on capacity required to meet a reliability standard established by Government. Final decision on capacity made by Government. Central auction to determine which providers (e.g. generation, DSR, storage) are issued with capacity contracts. All providers receive clearing price. Capacity providers may hedge their position in secondary markets Providers of capacity commit to being available when needed or face penalties in the delivery year. Costs of capacity shared between suppliers in proportion to their share of peak demand. Source: Based on DECC Electricity Market Reform: Capacity Market Detailed Design Proposals (June 2013) Capacity to procure The Secretary of State will decide on the target level of capacity in each auction based on meeting a reliability standard. The standard is based on a loss of load expectation, representing the acceptable number of hours that it is expected that supply will not meet demand. Consumers are unlikely to experience the shortfall in supply, with it more likely being managed through voltage control or controlled load shedding. The current view is a reliability standard of 3 hours would be appropriate, which is the same as France, but lower than Ireland (8 hours) and Netherlands (4 hours). National Grid will advise on how this reliability standard translates into the required level of capacity to procure in the auction. Their assessment will only be based on ensuring the reliability standard is met in the delivery year, and does not take into account capacity requirements in subsequent years. In order to allow a trade-off to be made between cost and security of supply in the auction a demand curve will be fixed in advance of the auction. For example, the Government may prefer to procure 1GW less than the target if the price of the most expensive GW is very high. There will also be a price cap to reduce gaming opportunities. The parameters of the demand curve will be set out in advance of the auction and cannot be altered once the auction begins. The current context

14 Frontier Economics January 2014 Auction The CRM is a market-wide mechanism and will be administered by National Grid as the TSO. It has been designed to be technology neutral i.e. it doesn t discriminate between which types of generating plant (or demand-side response) may be successful in the auction. However, there are notable exclusions from the list of eligible participants (see Table 1), such as low carbon generation and interconnected capacity. Low carbon technologies which are supported through existing low carbon support schemes are excluded due to concerns about overremuneration. Interconnected capacity is excluded because an acceptable means for its inclusion has not been found. Table 1. Eligible and ineligible capacity in the CRM auction Eligible capacity New plant Existing plant (including refurbishments) Demand-side response (DSR) Ineligible capacity Low carbon generation receiving support though CfDs, RO or Small-scale FITs Interconnectors (for first auction at least) Capacity <2MW unless aggregated Storage Permanent reductions in demand (EDR) Source: DECC The auction process begins with a pre-qualification stage which takes place at least four months prior to the auction. This involves mechanistic checks on the ability of the generator to deliver on their prospective capacity. The amount of capacity that each participant is able to bid into the auction will be determined by National Grid based on an assessment of the reliable capacity that each generator can be expected to provide. This de-rating of the plant s capacity would take account of the probability of unplanned outages, the length of scheduled outages and fuel availability. This is an important concept when considering the level of capacity that an interconnector should be allowed to be bid in to the CRM, which we discuss in section five. The main auction will then be held every year, four years prior to the delivery year. There will be a second auction one year ahead of delivery to fine-tune capacity and allow DSR an opportunity to bid into the CRM. The current context

January 2014 Frontier Economics 15 All generators bidding into the auction will receive the market clearing price, i.e. the price set by the most expensive successful bidder. However, in order to reduce the potential for existing generators to exploit market power to force up the market price, there will be restrictions on the bids existing generators are allowed to make. All pre-qualified capacity will be registered as either a price taker i.e. only able to bid up to a low price threshold, or a price maker, i.e. able to bid up to the price cap. Typically existing capacity will be price takers and new entrants (including DSR) price makers. The length of contract for capacity selected in the auction (i.e. an obligated party) will vary depending on the type of bidder. This is yet to be finalised but a new entrant could receive a ten year contract, and a major refurbishment three years. All other existing generation might receive a one year contract. Trading It is expected that obligated parties can trade their obligations in the market to manage the risk associated with not being able to fulfil their contracted obligations. For example there may be a number of scenarios where a participant may want to trade their obligation physically or financially: Maintenance generators will want to manage the risk of a market stress event during their planned maintenance. Plant economics change for example, cost increases making a previously agreed contract uneconomic, and creating a preference for trading out of the obligation. Delays in build time generators may wish to trade out of their obligation if they are experiencing severe delays in build and there would be more value in selling the contract. In response to one of these scenarios a party could choose to trade the obligation physically, at any point from one year ahead of delivery. National Grid will need to approve the trade based on the eligibility of the party receiving the contract. Plant with existing contracts are not allowed to increase them. However, trades could relate to unsuccessful plant in previous CRM auctions, new plant commissioning early, or other new capacity such as new DSR or de-mothballed plant. At any time parties can hedge their exposure to penalties financially through private agreements. The capacity provider is still liable for any penalties and will settle under the CRM as normal, recouping costs from the financial counterparty. The current context

16 Frontier Economics January 2014 Delivery The capacity contract obliges the participant to deliver energy at least to the correct proportion of their obligation in a market stress event. This event is defined as a period of voltage control or controlled load shedding for 15 minutes or more when at least four hours notice is provided. If the notice period is less than four hours then the delivery obligations are not binding. Successful delivery is based on a plant fulfilling the correct proportion of their contracted capacity. The obligations are load following so the proportion of the contracted obligation required is based on the proportion of all capacity required to meet total demand. Unsuccessful delivery during a stress event incurs a penalty which is based on the value of lost load minus the cash-out price 2, so generators face an economic incentive to deliver. The total penalty is capped in any one year, most likely at the total value of that year s capacity payments. Once the cap has been reached, the generators still have an incentive to be reliable because successful delivery can reduce their total exposure to penalties if they deliver in future stress events. Payment The total cost of the capacity market will be recovered from retail suppliers. These costs include the capacity payments and settlement body costs. Penalty payments received are netted off before the remaining costs are spread out over suppliers based on their share of peak demand. The peak demand is based on the methodology used for calculating triad charges. 3 2.2 The development of a CRM in the German market Unlike in GB, Germany has not yet made significant progress towards a decision on capacity markets. While in GB the government has reached the conclusion that the energy only market is insufficient to deliver security of supply, Germany is still unsure of its position on that key question. However they have considered a number of different options for the potential design of a CRM: Most focus has been around a centralised capacity market, which is the same high-level approach that has been adopted in GB. Within this market, a 2 The cash out price is the energy Imbalance Price under the British Electricity Trading Transmission Arrangements (BETTA). 3 The triad is a definition of peak demand used in National Grid s transmission charging methodology, based on maximum demand over three separate half hour periods with a specified minimum period between each half hour period. The current context

January 2014 Frontier Economics 17 regulated central authority such as a (transmission) system operator is responsible for the supply and settlement of the necessary secured capacity. The costs of the intervention are allocated administratively to the final consumer. The German Government has also considered alternative less interventionist models. A decentralised capacity obligation similar to the design proposed in France has also been considered. Under a decentralised system, market players (e.g. producers, retail suppliers or final consumers) are obliged to contract to secure capacity (kw) in a capacity market. This is in addition to their energy purchases in the energy market. Sufficient capacity can be realised by either self-supply or the purchase of certificates, which providers of secured capacity will receive. The obliged parties will try to pass-through the additional costs in their product prices to their customers. The current context

18 Frontier Economics January 2014 3 Impact of CRMs on market pricing We have already set out the European context for this work, where many Member States have introduced or are considering the introduction of CRMs. This section goes on to consider in more detail the potential impact of different CRMs on market pricing, and hence their impact on investment decisions for interconnectors. There are three broad types of CRM currently in operation or in planning across Europe, as set out in Table 2.Error! Reference source not found. Table 2. CRM options in operation and under consideration in Europe CRM option Quantity obligation (The options discussed in relation to GB and Germany fall into this category) Capacity payments Disturbance/ strategic reserve Source: Frontier Economics Description Authorities determine quantity of capacity required to serve demand based on a desired reliability standard Authorities arrange for purchase of this volume of capacity early enough to allow new build (central auction / decentralised obligation) Generators hence receive payments for capacity volumes and energy provided Penalties for those who contract to provide capacity and then fail Authorities determine a price to be paid for available capacity, on top of energy price Market design ensures this price is paid to generators available at a given time Penalties for those who receive payments and then fail Authorities determine a volume of capacity to be procured over and above that secured by market Authorities arrange for this volume to be purchased, with cost passed to customers This volume is excluded from the market other than in exceptional circumstances (e.g. when all upward regulation bids exhausted) The majority of the discussion in this report focuses on quantity obligations, because this is the model being implemented in the GB market and most under consideration in Germany. The principles and options presented later in the report could equally apply to a capacity payments model with only small modifications. But, the issues related to a strategic reserve are quite different. There is a brief discussion of the issues for interconnected capacity and strategic reserves later on in this section. Impact of CRMs on market pricing

Required revenues January 2014 Frontier Economics 19 3.1 Impact of CRMs on pricing Each of the different CRM options could be expected to have an impact on energy pricing, though to differing degrees. 3.1.1 Quantity obligation The impact of a quantity obligation is to provide a more certain and potentially larger revenue stream to the investor. Whilst in theory the CRM can restore the missing money for investors, it could also reduce the share of revenues from the wholesale market through lower wholesale prices, as illustrated by Figure 3. Figure 3. Capacity market has the potential to reduce wholesale market revenues Potential missing money leading to underinvestment Capacity market revenue Wholesale market revenue Wholesale market revenue Energy only market CRM Source: Frontier Economics A credible capacity mechanism would result in an expectation that the defined reliability standard would be met. This would in turn result in lower expectations of scarcity rents which, depending on how competition in the market plays out between generators, may lead to more competition at any given demand level, and hence prices which are closer to SRMC. The capacity remuneration will then fill the gap between revenues based more closely on SRMC and the cost of investment. While prices are likely to fall, they may not fall all the way to SRMC. The extent of mark-up over SRMC is likely to depend on the level of competition present in the market. Even if competition with the CRM is still relatively weak at any given demand level and there are opportunities to extract scarcity rents, then the threat of regulatory intervention could still result in prices closer to SRMC. This is because whilst scarcity rents may be tolerated by the regulator in an energy only market, they are in principle no longer required to incentivise investment given the Impact of CRMs on market pricing

20 Frontier Economics January 2014 presence of the CRM. Pricing above marginal cost is therefore more likely to be viewed as an abuse of market power. If wholesale market revenues do not fall, this may initially lead to overremuneration of capacity. After a time, as participants understand the impact of the CRM better on wholesale prices, provided there is freedom of entry the clearing price in the CRM auction should fall to re-dress the balance between capacity and wholesale market revenues. The energy price should then in turn fall. In other words, if the generation market is competitive and there is freedom of entry, the combined effect of the energy and capacity markets is likely to be to reduce average energy prices and also to reduce energy price spikes above SRMC. 3.1.2 Capacity payments The logic is very similar for a capacity payments regime. Provided the payments, which are set ex ante, are set at the right level they should incentivise new entry by filling the gap between the revenue required by a new entrant and the expected market revenues. The expected effect on energy market revenues will be the same as that discussed for a quantity obligation, i.e. energy prices may be closer to SRMC than without the capacity payment. 3.1.3 Strategic/disturbance reserve A strategic reserve may also reduce the energy price in the market, but to a much lesser extent than for the capacity obligation or capacity payment. Typically, a strategic reserve is a volume of generation which is deliberately withheld from the market except in circumstances where involuntary load shedding would otherwise take place. It is often priced at an estimate of the value of lost load (VoLL). 4 Therefore, if its governance is sound, a strategic reserve should not affect normal market pricing outside of very low probability events when it might be called into use. It should only set the price when there is no other source of voluntary supply available. However, during these low probability events, the reserve will be setting the price based on an estimate of VoLL. This is an administered price that could underestimate the true value of VoLL and therefore reduce revenues (inframarginal rents) to all other market participants. To the extent that the strategic reserve does reduce some very peak prices during low probability events, it could weaken the signal for new investment, increasing the future quantity requirement for the strategic reserve. 4 Sometimes the strategic reserve is priced between VoLL and the highest SRMC of plant or DSR on the system. This would have implications for market prices when the strategic reserve is called and hence for generation revenues. Impact of CRMs on market pricing

January 2014 Frontier Economics 21 The nature of plant selected to act as a strategic reserve can also affect the level of price distortion. For example, if relatively inflexible plant act as strategic reserve, they may need to be brought onto the system early (i.e. while there are still other plant capable of meeting demand that are not running at full output) in order to ensure that they are at full output when demand peaks. During these ramp up periods, revenues for other peaking plant would be negatively affected by the presence of the strategic reserve plant. 3.2 Impact of CRMs on interconnectors For generators within the CRM, a shift to energy pricing based more closely on SRMC may not be a concern. The gap between expected market revenues and the required revenue for new entry in the market should be filled by the capacity market revenue. However, interconnected capacity, which is outside the CRM, faces the potential of a reduced market price while not benefitting from the additional capacity market revenues 5. The impact of the lost mark-up over the SRMC is potentially quite significant. Evidence from the GB and German markets illustrates how peaky energy prices have been historically, as shown in Figure 4. 5 The strategic reserve option is slightly different in that all generators outside of the reserve, domestic or interconnected, are affected by the change in the price, and therefore could potentially lose revenues. Impact of CRMs on market pricing

22 Frontier Economics January 2014 Figure 4. Observed prices without a CRM in the GB, German and Norwegian markets 180 First two weeks November 2012 160 140 120 100 80 60 40 20 0-20 GB day-ahead (APX-N2EX weighted) DE day-ahead (EPEX spot) NO day-ahead (Nordpool,Kristiansand) Source: Frontier Economics The lost revenue to interconnected capacity will have knock-on impacts for investment in interconnectors. For example, interconnector owners receive a congestion rent based on the expected spread between prices on either end of the interconnector. If the spread is reduced, then this feeds through to the revenues of the interconnector owner and weakens the case for investment. In the period shown in Figure 4, reducing peak prices in GB and Germany would narrow the price spread with Norway and hence the profitability of a potential interconnector. This is illustrated in a simulation for 2025 based on a new link between Norway and the GB market. A CRM in the GB market could reduce the price spread by around 10%. Impact of CRMs on market pricing

January 2014 Frontier Economics 23 Figure 5. Simulation of new link between GB and Norway in 2025 Average GB price in this forecast is 81 (2012 real), mainly driven by assumptions on gas and carbon prices. Average GB price in 2012 was 45 Source: Frontier Economics A CRM on both sides of the interconnector may maintain price spreads if the CRMs had the same effect on energy market prices in the two areas. However, even then excluding interconnectors from the CRMs could result in under investment in interconnector capacity relative to domestic generation capacity since the domestic generators would receive revenues from the CRM and the interconnector would not. Impact of CRMs on market pricing

24 Frontier Economics January 2014 4 Arguments for the participation of interconnected capacity in CRMs CRMs have the potential to have significant impacts on interconnection. This section discusses in more detail the arguments for their inclusion. Our consideration of whether interconnected capacity should be included in a CRM rests on two sets of high-level questions: Security of supply: Is an interconnector physically as reliable as domestic generation? Will an interconnector always deliver power when it is required? Economic efficiency: How will excluding interconnectors impact on the operational efficiency of the electricity market? How will investment and closure decisions be affected? We describe each of these questions in turn, and then attempt to assess the potential efficiency loss of not including interconnected capacity in a CRM. Finally, we consider how the inclusion or not of interconnected capacity could be viewed by the European Commission, and in particular how this could be viewed from a State Aid perspective. 4.1 Security of supply Electricity is a homogeneous product, so there is no difference in the power delivered to the consumer by an interconnector or a domestic generator. Interconnectors are historically physically reliable (at least to the same level as a domestic generator). And, even if this were not the case, the interconnector s capacity could be adjusted for reliability. We have already noted that in the proposed design of the GB CRM, the amount of capacity each generator is allowed to bid into the CRM is adjusted for its physical reliability based on historic data. However there are two potential areas of concern. First, shocks may occur coincidentally between balancing zones at either end of the interconnector, and second, markets may not be perfect and so power may not flow to the market with the highest price (which should be the market with the greatest degree of stress ) Arguments for the participation of interconnected capacity in CRMs

January 2014 Frontier Economics 25 4.1.1 Coincident stress events If the markets on both sides of the interconnector are experiencing scarcity, then it is not certain in which direction the power will flow. In theory prices should rise in each market with power ultimately flowing towards the market willing to pay more for power instead of experiencing demand reduction (i.e. the market with the greatest degree of scarcity or the higher VoLL). However, the market is unlikely to operate in this way without an active demand-side. A possible outcome is that no power flows over the interconnector as each TSO takes action to prevent exports at a time of stress. The design of CRMs will need to take into account the potential for such coincident stress events. In the context of a GB interconnector with Norway, this probability is likely to be lower than with more correlated Central West European markets. 4.1.2 Imperfect markets Power flows over interconnectors do not always flow towards the country with the highest price, suggesting that there are inefficiencies in their dispatch of power. Therefore, the interconnector may not be able to deliver power at times of stress. The existence or severity of a stress event may not always be known a day ahead of delivery, so the outcome of day ahead market coupling will not be able to take it into account. In which case, the ability of an interconnector to fulfil its obligation rests at least in part on the efficiency of the intra-day market and cross border balancing arrangements between TSOs. Resolving this issue is clearly of critical importance however, it is related more to the design of energy markets than the CRM itself. We return to it below in the context of changes to the energy market which should accompany any design of CRM. 4.2 Economic Efficiency The discussion of economic efficiency focuses on whether excluding interconnectors from CRMs will have an impact on operational or investment efficiency. 4.2.1 Operational efficiency The impact on operational efficiency depends on the impact that the CRM has on prices and hence on operating decisions. In section three we set out some of the reasons why it is possible that a CRM could lead to lower energy market prices, noting that the scale of the impact was uncertain. Arguments for the participation of interconnected capacity in CRMs

Price, Costs Price, Costs 26 Frontier Economics January 2014 We therefore discuss briefly two scenarios. First, one in which energy prices are unaffected by the CRM and second, one where energy prices are reduced as a result of the CRM i.e. they are more reflective of SRMC. Prices unaffected by CRM In this scenario, the price-setting behaviour for both domestic and interconnected capacity is based on SRMC 6 plus a mark-up. As long as the price mark-up behaviour is consistent in both markets then generation across the combination of the two markets will be dispatched according to the lowest SRMC (consistent with the available interconnector capacity), and hence there will be no distortion to operating signals a result of the CRM. Figure 6. No distortion to trade over the interconnector Country A (CRM) Demand + net exports Country B (No CRM) Demand + net exports P* SRMC + mark-up SRMC + mark-up MWh MWh Source: Frontier Economics This result does not change if there is also a CRM in the neighbouring market. Prices lower as a result of the CRM In this scenario, the CRM is implemented in the local market and not in the interconnected market. Wholesale prices are now more reflective of SRMC in the market which has the CRM. Interconnected generators do not receive any capacity payments so they will be bidding based on their SRMC plus a mark-up to cover their fixed costs of investment. Therefore a domestic generator with a higher SRMC would be 6 The SRMC is the variable cost of an additional unit of power, the majority of which is the cost of fuel. In the Norwegian hydro-based system where a significant amount of energy is stored in reservoirs the direct cost of fuel is zero. However, there is an opportunity cost associated with generating which generators factor into their scheduling decisions. This cost i.e. the revenue that could have been earned from storing the water for future use, forms the key part of the SRMC in the Norwegian system. Arguments for the participation of interconnected capacity in CRMs

Price, Costs Price, Costs January 2014 Frontier Economics 27 chosen in preference to an interconnected generator with a lower SRMC, as long as the difference in SRMC was less than the mark-up. This leads to a welfare loss. This scenario is shown in Figure 7. The pricing in Country A with a CRM is based on SRMC. Generators in Country A displace generators in Country B which have a lower SRMC but still include a mark-up. Overall the cost of generation across the two countries has increased with the introduction of the CRM in the domestic market, leading to a welfare loss due to higher operating costs. Figure 7. Distortion to trade over the interconnector leading to a welfare loss Country A (CRM) Country B (No CRM) Demand + net exports Demand + net exports P* SRMC + mark-up SRMC + mark-up P** SRMC P** Welfare loss MWh MWh Source: Frontier Economics In a scenario where there is also a CRM in the neighbouring market, provided the impact on the price is uniform in both markets the potential distortion is removed with generation dispatched according to their SRMC across the two markets. However, given the range of different CRM designs being considered, the impact on price is unlikely to be uniform across markets, suggesting that a distortion could remain. Impact of including interconnectors in CRM on operational efficiency There is the potential for a distortion in short-run prices and hence trade over the interconnector when a CRM is introduced. However, including interconnected capacity in the CRM may not lead to a significant improvement. This is because some or all of the value of any capacity payments is likely to flow through to the interconnector owner, directly or indirectly. If the interconnector owner participates directly in the CRM itself then they will receive the full value of the capacity payments. If interconnected generators participate in the CRM, much of this value will be passed through to the interconnector owner when bidding for the right to participate in the CRM. As a result the bidding behaviour of generators (in relation to the energy market) remains largely unchanged whether or not they can participate in the CRM. Arguments for the participation of interconnected capacity in CRMs