Executive Summary Since February 2009, Ofgem has produced a series of reports examining margins earned by energy suppliers in Great Britain. These reports provide a forward-looking measure of revenues, costs and margins over the coming year, which Ofgem currently refers to as the Supply Market Indicator (SMI). 1 Ofgem’s methodology clearly states that the “SMI cannot and does not seek to provide a forecast of company profits” and that profitability can only be assessed retrospectively.2 Separately, Ofgem has stated that the historical Consolidated Segmental Statements (CSS) produced by large energy suppliers provide a “reliable indicator of profitability”.3 As an empirical matter, the various incarnations of the SMI have repeatedly overstated profits in gas and electricity supply compared to historical information released in the Consolidated Segmental Statements (CSS).4 Over the period since 2009 for which the CSS have been available, the forward-looking margins in the SMI have exceeded the backward-looking margins in the CSS by 50-200 per cent on average (equivalent to £29 per electricity-only, £22 per gas-only and £17 per dual fuel customer per year). The most recent incarnation of Ofgem’s SMI, which uses an updated method, overstates profits by 18 to 100 per cent compared to the latest available segmental statements from 2013. The most obvious practical reason for differences between the SMI and the CSS is that the SMI is forward-looking. The actual course of events, such as changes in the weather or wholesale costs, is difficult to predict. However, the SMI’s forward-looking nature alone does not explain what appears to be a systematic upward bias in the profit levels estimated. Energy UK commissioned NERA to conduct an independent, high-level review of the SMI methodology published by Ofgem, and to identify potential systematic biases that explain the persistent overstatement of profit levels using publicly available data. Our review identified simplifications in Ofgem’s approach that could artificially increase or decrease its estimated margins. We also identified factors which may bias the SMI estimate of margins upwards, including but not limited to the following: Ofgem’s assumed level of consumption may overstate the consumption of domestic households (and is higher than figures that Ofgem states are “typical” in other documents); 5 Ofgem appears to take a simple average of the fixed tariff offered by each supplier and weights these six tariffs by market share as part of its calculation of average annual bills. 6 1 Ofgem (2009), Quarterly Wholesale/Retail Price Report, ref 15/09, February 2009. 2 Ofgem (2014), Supply Market Indicator – Methodology, 27 November 2014, page 3, para. 2.3. 3 Ofgem (2014), page 3, para. 2.3. 4 Both the SMI and CSS calculate measures of pre-tax profit, a net pre-tax margin. 5 Ofgem calculates “Tariff Comparison Rates” using “typical” values of domestic consumption of 3,200kWh of electricity and 13,500kWh of gas, below the 3,800kWh and 14,600kWH assumed in the SMI for 2015.Energy Supply Margins: Commentary on Ofgem's SMI Executive Summary NERA Economic Consulting ii In practice, customers may choose proportionately more tariffs towards the bottom end of the available range and therefore pay less than the simple average; Ofgem’s approach to calculating wholesale costs reflects information gathered in 2008 and may no longer be appropriate to current market conditions. Moreover, Ofgem’s estimated wholesale cost does not appear to include the full costs of shaping wholesale purchases to the profile of customers’ demand; and Ofgem assumes that indirect costs per customer will remain broadly flat in real terms based on evidence that total indirect costs have been broadly flat over time. In practice, if suppliers have fixed costs in the short run, recent changes in market share for the larger suppliers could have caused costs to rise on a per-customer basis. Our preliminary analysis suggests that, with minor departures from existing assumptions to adjust for some of these biases, the latest SMI could overstate profits by around £35 per electricity account, £19 per gas account and £54 per dual fuel account. These differences account over 100 per cent of the average margin on electricity and dual fuel customers in the 2013 CSS and 67 per cent of the margin on gas customers, i.e., in the same range as the historically observed bias in the SMI. 6 Ofgem states that the average is weighted for supplier market shares, but not for the prevalence of each tariff for each supplier “The annual customer bills across all tariffs are calculated based on annual domestic consumption. We then calculate the average annual customer bill weighted by suppliers’ market shares and payment method.” Ofgem (2014), Supply Market Indicator Methodology, 27 November 2014, page 8, para 3.8.Energy Supply Margins: Commentary on Ofgem's SMI Introduction NERA Economic Consulting 1 1. Introduction Energy UK Commissioned NERA to review the methodology for Ofgem’s Supply Market Indicator (SMI). Our report provides an independent assessment of Ofgem’s analysis. Since February 2009, Ofgem has produced a series of reports examining margins earned by energy suppliers in Great Britain.7 The initial method relied on gross margins and relied on a number of assumptions which did not accurately reflect suppliers’ revenues and costs. In late 2009 and for a series of subsequent reports, Energy UK commissioned NERA to calculate alternative estimates using more realistic assumptions.8 Ofgem’s methodology has changed significantly over time. This brief report reviews Ofgem’s latest methodology and identifies potential biases in Ofgem’s approach. Although a detailed quantification and estimate of margins is beyond the scope of this report, where possible, we have identified the possible magnitude of these biases. We also highlight the uncertainty around Ofgem’s estimates by pointing to cases where Ofgem uses unreliable data or approaches which may bias its estimates of net margins in either direction. The report proceeds as follows: Section 2 sets out the purpose of Ofgem’s SMI and calculates its historical bias relative to the CSS; and Appendix A describes each of the major revenue and cost components that form the SMI, and highlights the uncertainty and any potential bias in Ofgem’s method. Unlike in previous reports, we have not relied on any proprietary data submitted to us by Energy UK or the large energy suppliers to conduct our analysis, and where we have quantified potential biases in Ofgem’s methodology we have relied on publicly available information. 7 Ofgem (2009), Quarterly Wholesale/Retail Price Report, ref 15/09, February 2009. 8 NERA (2009), Energy Supply Margins At Current Prices, 7 December 2009, page 13.Energy Supply Margins: Commentary on Ofgem's SMI Purpose of Ofgem’s SMI NERA Economic Consulting 2 2. Purpose of Ofgem’s SMI 2.1. Ofgem Has Updated Its Method of Calculating the SMI over Time Since February 2009, Ofgem has produced a series of reports examining margins earned by energy suppliers in Great Britain.9 Initially, Ofgem released the reports on an approximately quarterly basis and focussed on the difference between the direct costs incurred by energy supply businesses and retail tariffs (the “gross” margin).10 In late 2009 Energy UK commissioned NERA to calculate an equivalent “net” margin figure for supply businesses to remove any misconception that Ofgem’s calculations represented profits earned by suppliers. After discussion with NERA, Ofgem’s quarterly reports on the retail energy market presented both gross and net margins from February 2010 onwards.11 NERA’s and Ofgem’s publications differed in some important respects. Notably, Ofgem modelled the margins on the basis of a representative customer on a standard variable tariff for an assumed level of consumption and based on cost estimates from the Probe. NERA’s calculations, based on a survey of the suppliers, incorporated a wider range of tariffs, average consumption levels and updated cost estimates. The difference between NERA’s and Ofgem’s estimates of margins varied between rounds but accounted for around £65 per dual fuel customer by March 2011.12 By January 2012, Ofgem began publication on a weekly basis and was referring to its margin estimates as the “Supply Market Indicator” (SMI). Since then, Ofgem has made a number of changes to its methodology for estimating margins to incorporate more realistic assumptions on the revenues earned and the costs incurred by energy suppliers. As part of the latest round of changes, Ofgem moved to monthly publication in March 2014.13 Despite the changes in the methodology designed to adopt more realistic assumptions, Ofgem makes clear that “the SMI cannot and does not seek to provide a forecast of company profits”.14 The SMI relies on historical assumptions on costs, consumption and hedging strategies and compares them with current (or announced) price levels. The comparison of historical costs with current prices means that the SMI may not be an accurate indicator of out-turn margins, even on average and over the long term. Ofgem readily recognises the inadequacy of the SMI as a measure of industry profit. In the methodology for the SMI, Ofgem argues that profitability can only be observed retrospectively with reference to historical data. 15 Elsewhere, Ofgem acknowledges that “the 9 Ofgem (2009), Quarterly Wholesale/Retail Price Report, ref 15/09, February 2009. 10 Ofgem (2009), Quarterly Wholesale/Retail Price Report, ref 15/09, February 2009, section 5, page 11 of 12. 11 Ofgem (2010), Electricity and Gas Supply Market Report, Ref 23/10, 22 February 2010. 12 NERA (2011), Energy Supply Margins Estimated by Ofgem and NERA: Review of the Differences –Commissioned by Energy UK, 20 October 2011, page 1. 13 Ofgem (2014), Supply Market Indicator Methodology, 27 November 2014, page 6, para 2.12. 14 Ofgem (2014), Supply Market Indicator – Methodology, 27 November 2014, page 3, para. 2.3. 15 “Only by looking back can we accurately observe a company’s profitability in a given time period.” Ofgem (2014), Supply Market Indicator Methodology, 27 November 2014, page 3, para 2.3.Energy Supply Margins: Commentary on Ofgem's SMI Purpose of Ofgem’s SMI NERA Economic Consulting 3 SMI should be viewed as an estimate rather than a factual report” and that its purpose is to “help consumers and other stakeholders understand pricing trends” rather than identify profit levels.16 2.1. The SMI Has Overstated Profits Historically Ofgem notes that “a reliable indicator of profitability (before interest and taxes) is provided in the individual companies’ Consolidated Segmental Statements (CSS)”.17 Following the publication of annual CSS, Ofgem presents the figures on a per customer basis in a manner that can be directly compared with the SMI. Prior to March 2014, and the latest significant revision to Ofgem’s method, this comparison shows that the SMI tended to systematically overstate the actual profitability of the large suppliers. Table 2.1 presents Ofgem’s 12-month forward-looking “indicative net margin” alongside the actual backward-looking net margin from the CSS. For instance, the row entitled “Jan-09” shows the forward looking net margin calculated using Ofgem’s SMI model for the twelve months of 2009, against the contemporaneous net margins actually earned by suppliers and published by Ofgem as “historic actuals” taken from the CSS. As can be seen from Table 2.1, in the period January-December 2009, Ofgem’s SMI understated actual profit margins on dual fuel customers (by £30 per customer). However, in all other periods (and in every period for single fuels) Ofgem’s method overstated profit margins. Over the whole period therefore, Ofgem overstated margins on average by £17 per dual fuel customer, £22 per electricity customer and £29 per gas customer. In other words, Ofgem’s method gave the impression that large suppliers’ profits per customer were 50-200% higher than actually revealed by the suppliers’ published CSS, which Ofgem argues is a “reliable indicator of profitability”. 16 Ofgem (2014), page 3, para. 2.2. 17 Ofgem (2014), page 3, para. 2.3.Energy Supply Margins: Commentary on Ofgem's SMI Purpose of Ofgem’s SMI NERA Economic Consulting 4 Table 2.1 The SMI Has Persistently Overstated Profit Margins, by £17-29 per Customer: 12 Month Net Margin (£ per Customer) Source: NERA analysis of Ofgem data. “Estimated” numbers are Jan-Dec SMI estimates published by Ofgem under its old methodology.18 “Actual” numbers are derived by Ofgem on a comparable basis from companies’ CSS.19 For comparability, numbers presented here are not rolling averages. No forecast will ever provide a completely accurate view of the future, and Ofgem notes that its SMI is based on “several factors that are inherently uncertain … [and] should be interpreted with this in mind”.20 However, an unbiased forecast should lead to both positive and negative errors over time. Table 2.1 demonstrates that Ofgem’s SMI has persistently overstated profitability, which may suggest that the method used to derive it is inherently biased for the purpose of forecasting profits or that it would only be a reliable estimate over the (very) long term. 2.2. The Updated Method Appears to Suffer From a Similar Bias Ofgem made major revisions to the method it uses to derive the SMI in March 2014. Among other changes, Ofgem revised its presentation of costs and revenues so that its forwardlooking estimates can be compared with the historical information contained in the CSS on a line-for-line basis. Ofgem substitutes the data on revenues, costs and margins calculated by its SMI model with those from the CSS for historical periods, as the CSS data becomes available. For January-December 2013, the first full year in which comparable information is available, the SMI still appears to provide a biased view of profitability. Table 2.2 compares Ofgem’s estimates for January to December 2013 with the cost information recorded in the CSS for electricity-only, gas-only and dual-fuel customers. Compared to the CSS, the SMI overstated the average revenue per customer (by £4 for electricity accounts and £8 for gas accounts) while underestimating the average wholesale cost per customer (by £5 per electricity account, £15 per gas account and £21 per dual fuel account). Other differences had a very small net effect (£1 or less per account). Differences in the revenue line between the SMI and CSS hinder direct comparison of the margins of single and dual-fuel customers. The CSS breaks costs, revenues and margins 18 Ofgem (2013), Electricity and Gas Supply Market Indicators, 10 January 2013. 19 Ofgem (2014), Supply Market Indicator for November 2014, 27 November 2014. 20 Ofgem (2014), Supply Market Indicator Methodology, 27 November 2014, page 3, para. 2.2. Gas Electricity Dual Fuel Actual Estimate Diff. Actual Estimate Diff. Actual Estimate Diff. Jan-09 -£2 £0 -£2 £12 £15 -£3 £10 -£20 £30 Jan-10 £39 £75 -£36 £2 £25 -£23 £41 £75 -£34 Jan-11 £25 £50 -£25 £8 £30 -£22 £32 £50 -£18 Jan-12 £47 £65 -£18 £10 £35 -£25 £56 £65 -£9 Jan-13 £29 £90 -£61 £20 £55 -£35 £49 £105 -£56 Average £27 £56 -£29 £10 £32 -£22 £38 £55 -£17Energy Supply Margins: Commentary on Ofgem's SMI Purpose of Ofgem’s SMI NERA Economic Consulting 5 down into gas accounts and electricity accounts, whilst the SMI attributes different revenues to gas, electricity and dual fuel customers. Specifically, in its estimated SMI, Ofgem assumes that the total costs for an average dual fuel customer are equal to the sum of the costs for an average electricity and gas customer, but the revenue a supplier receives for a dual fuel customer is £24 (1.8 per cent) lower. As a result, the margin on a dual fuel customer derived from the CSS is likely to overstate the true margin suppliers earn, as it excludes discounts (although there will be a corresponding understatement of the margin earned on single fuel customers). Overall, the differences noted above result in a difference in pre-tax margin of £9 to £23 per customer account per year. These are sizeable when compared to the actual margin per customer that suppliers achieved in 2013 of £29, £20 and £49 per customer respectively for gas, electricity or dual fuel. In other words, in 2013, the SMI estimates of margins ranged between 18% and 115% higher than the actual margins reported in the CSS. Table 2.2 The Updated SMI Continues to Overstate Revenues and Understate Costs: Jan-Dec 2013 (£ per Customer, 2014 prices) Source: NERA analysis of Ofgem data. “Estimated” numbers are the Jan-Dec 2013 SMI estimates published by Ofgem in July 2014, while “Actual” are CSS numbers presented by Ofgem on a comparable basis in August 2014. N.b. for comparability between CSS and SMI data, we do not present margins on a rolling average basis. 2.3. Explaining the Inherent Bias in Ofgem’s SMI In Appendix A, we examine in the method Ofgem uses to estimate each major line item that makes up the SMI. We provide a (non-exhaustive) list of possible biases and uncertainties in Ofgem’s method, which may explain the persistent overstatement of costs we observed in sections 2.1 and 2.2, above. We also provide some illustrative examples to quantify the approximate range of the effects. We summarize our major findings below: Revenue: Ofgem’s revenue assumptions may be biased by its simplified estimates of consumption and its approach to incorporating fixed tariffs. A simple sensitivity that we ran for the second effect suggests that the bias may lie in the region of £5-£8 per customer account (roughly one per cent of a dual fuel bill). Gas Electricity Dual Fuel Actual Estimate Diff. Actual Estimate Diff. Actual Estimate Diff. Revenue £734 £738 -£4 £582 £590 -£8 £1,316 £1,304 £12 minus Total Costs £705 £700 £5 £562 £547 £15 £1,267 £1,246 £21 of which; Wholesale Costs £396 £390 £6 £247 £231 £16 £643 £621 £22 Netw ork Costs £144 £148 -£4 £138 £140 -£2 £282 £288 -£6 Environmental & Social Obligations £38 £33 £4 £74 £72 £2 £111 £105 £6 Other Operating Costs £92 £93 -£1 £76 £76 £0 £168 £169 -£2 VAT £35 £35 £0 £28 £28 £0 £63 £62 £1 equals Pre-Tax Margin £29 £38 -£9 £20 £43 -£23 £49 £59 -£9Energy Supply Margins: Commentary on Ofgem's SMI Purpose of Ofgem’s SMI NERA Economic Consulting 6 Wholesale Costs: Changes in wholesale prices may take longer to feed in to reductions in the costs borne by suppliers if suppliers procure further ahead than Ofgem assumes. In practice that could account for suppliers’ costs being £3-7 higher per account (roughly one per cent of a dual fuel bill). Similarly, Ofgem’s analysis appears to exclude the costs of matching the profile of wholesale energy to the profile of customer demand (the cost of “shape”), which could cost around a further £20 per electricity customer and £5 per gas customer (three per cent and one per cent of electricity and gas bills respectively). Indirect Operating Costs: Ofgem assumes that unit operating costs per customer rise in line with inflation. In practice, if suppliers’ costs are fixed in the short term, an increase in the market share of entrant suppliers may cause existing suppliers’ costs to rise by up to a further £3 per customer account per year (less than one percent of a dual fuel bill). Network Costs: Ofgem’s method for calculating network costs is dependent on its assumed level of consumption, which ignores the five million customers that receive electricity through Economy 7 meters, and consume 60 per cent more electricity than standard meter customers. This may understate the network costs suppliers face to meet the demand of an average electricity customer, although we have not quantified this effect. In addition, the weighting of local distribution and transmission tariffs that Ofgem uses that may not accurately reflect the actual location of domestic demand, and could lead to inaccuracies in the SMI that could in principle be positive or negative. Environmental and Social Obligations: Ofgem has assumed that the cost to suppliers of delivering the Energy Company Obligation will remain flat over time, ignoring published estimates from DECC that suggest that the cost of this obligation is likely to increase in the next 12 months. Since suppliers are not yet close to meeting their targets for energy efficiency measures under the policy, there appears to be a one-sided risk that this cost will increase rather than decrease. This is another potential area of bias in the SMI that is likely to lead to an overstatement of future profits. We quantified the possible bias in the revenue calculation, wholesale costs and indirect operating costs using simple illustrative examples. In combination, those estimates would suggest that the SMI overstates profits by around £35 per electricity customer account, £19 per gas account and £54 per dual fuel account. Those margins account for 6 per cent, 3 per cent and 4 per cent of a customer bill respectively or over 100 per cent of the average margin on electricity and dual fuel customers in the 2013 CSS and 67 per cent of the margin on gas customers. 2.4. Possible Biases May Lead to Overstatement of Profits Our analysis above illustrates that the SMI has consistently overstated profits relative to the Consolidated Segmental Statements (CSS), which Ofgem argues are reliable measures of profitability. Although Ofgem has repeatedly adjusted its method for assessing margins, the most recent SMI Methodology still appears to overstate margins relative to the CSS. We have identified possible biases as well as simplifications with uncertain effects in Ofgem’s current methodology. These biases and simplifications may account for some or all of the tendency of the SMI to overstate the profitability of large suppliers. Our simple, illustrative adjustments to Ofgem’s method show that relatively minor changes to the Energy Supply Margins: Commentary on Ofgem's SMI Purpose of Ofgem’s SMI NERA Economic Consulting 7 underlying assumptions to correct for identifiable biases in its method cause margins under the SMI to fall substantially.Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 8 Appendix A. Components of the SMI This Appendix reviews each of the key assumptions that feed into Ofgem’s SMI model, describes Ofgem’s approach briefly and offers insights into the inherent biases in the method. In some cases we have briefly investigated the potential magnitude of the effects but these calculations should be seen as preliminary, illustrative assessments rather than detailed examinations. The sections below appraise Ofgem’s approach to the following determinants of margins in the SMI: Section A.1: Consumption; Section A.2: Revenue; Section A.3: Wholesale costs; Section A.4: Indirect/operating costs; and Section A.5: Network charges and social obligations. A.1. Consumption Consumption figures feed into the estimates of revenues and costs and are therefore important in assessing the overall level of bills. Ofgem applies a simple, uniform consumption assumption in its SMI model. In this section we discuss Ofgem’s assumptions. We explain the potential biases it causes in the estimation of revenue in section A.2 below. A.1.1. Ofgem takes average domestic consumption figures from DECC Domestic consumption of energy has been trending downwards since Ofgem began its SMI exercise in 2009. To account for this, Ofgem states in its SMI Methodology that it uses the “latest average mean weather-corrected consumption values” available from DECC for 2013, to which it applies a demand forecast to arrive at 2014 and 2015 forecasts of mean consumption.21 The base year values estimated by DECC are: Gas: 14,800 kWh; and Electricity: 3,800 kWh. In the case of electricity, Ofgem uses an average value for customers on unrestricted meters (i.e. excluding customers using Economy 7 meters). Consumption by customers with Economy 7 meters is significantly higher, and is estimated by DECC as 6,000kWh.22 In NERA’s previous work for Energy UK, we conducted surveys of the large energy suppliers to identify the actual consumption of domestic customers. In the last such survey we conducted, in January 2012, we identified that the average annual consumption of an electricity customer on a standard meter was 3,974 kWh (slightly above the current value 21 Ofgem (2014), Supply Market Indicator – Methodology, 27 November 2014, para. 3.4, page 8. Ofgem applies a demand forecast to the base year values it takes from DECC. 22 DECC (2014), Special feature – Revisions to DECC domestic energy bill estimates, March 2014, Table 2, page 53. Consumption is rounded to the nearest 100 kWh.Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 9 used by Ofgem), while a gas customer consumed 14,410 kWh (lower than Ofgem’s value).23 If Ofgem is correct to assert that demand has fallen over time, it would follow that NERA’s survey would return yet lower consumption for gas and electricity today. In an earlier report, we identified the reason that Ofgem’s figure for average gas consumption is likely to overstate the true consumption of a domestic customer.24 Meters classified as domestic customers include in practice a large, but unknown, number of meters for small businesses with annual consumption below 73,200kWh (much higher than household levels). The inclusion of these heavier, non-domestic users in the data will tend to bias the estimated levels of consumption upwards, and hence overstate the volumes sold by suppliers and the revenue earned. A.1.2. Ofgem applies a forecast of change in total demand to household demand Ofgem’s accounts for falling domestic demand for gas and electricity by applying a declining aggregate demand forecast taken from National Grid to its assumed levels of household consumption. 25 If domestic customer numbers are stable over time, then consumption per household will indeed fall at the same rate as total consumption. However, the most recent data available from DECC indicates that the number of domestic customers is growing over time.26 Applying an aggregate demand forecast to the historical level of household consumption will understate the true decline in domestic demand, and may consequently overstate future domestic consumption. A.1.3. Ofgem elsewhere uses lower consumption numbers for “typical” households In separate publications, Ofgem calculates the dual fuel bill of a “typical” household.27 These “typical” values are: Gas: 13,500 kWh; and Electricity: 3,200 kWh. It is not clear how a “typical” household differs from an “average” household in this context or which is the appropriate standard for the SMI. However, Ofgem uses these levels of consumption to inform consumers about potential bill savings, suggesting that it believes them to be representative and informative. 23 NERA (2012), Energy Supply Margins: Update January 2012, 23 January 2012. 24 NERA (2011), Quarterly Report on Retail Electricity and Gas Markets in Britain: a Retrospective, 31 May 2011, pages 7-8. 25 Ofgem (2014), Supply Market Indicator – Methodology, 27 November 2014, Appendix 1. Ofgem uses National Grid’s “Gone Green” forecasts from its Future Energy Scenarios. 26 DECC (2014), Quarterly domestic energy switching statistics, 18 December 2014, Electricity customers increased by 217,000 in the year to September 2014 (0.8 per cent) while the number of gas customers rose 200,000 (0.9 per cent). 27 See, for example, Ofgem (2015), Energy customers who don’t fix could be paying up to 250 more than needed, 23 January 2015. Ofgem uses these consumption figures to calculate “Tariff Comparison Rates”.Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 10 A.2. Revenue Ofgem’s revenue calculations may not accurately reflect the bills recovered from customers. In this section we briefly discuss two possible sources of upward bias in its method: (1) Ofgem’s consumption assumption, and (2) its approach to incorporating fixed tariffs in its model. A.2.1. Ofgem’s consumption assumptions may bias revenues upwards In principle the impact of reducing consumption levels in the SMI model could increase or decrease the calculated margins, depending on the relationship between suppliers’ prevailing tariffs and the balance of their fixed and variable costs. We attempted to quantify the difference in the bill between an average customer, using Ofgem’s SMI consumption figures, and a typical customer, using the lower figures Ofgem uses for Tariff Comparison Rates. To do this we compared both the cheapest variable price tariff available from the large energy suppliers and the cheapest fixed price tariff (both for an example postcode). For each, we calculated the annual bill using Ofgem’s “typical” consumption values and for the consumption values used in the SMI. Table A.1 shows that this amounted to a difference in bills of between £102-£131. This suggests that Ofgem’s choice of a higher consumption level for calculating the SMI overstates the revenue earned per customer (relative to a “typical” household) and, assuming not all cost are variable, overstates the net margin per customer. Table A.1 "Typical" Consumers Pay Much Lower Bills Than The "Average" Consumer Used in the SMI Source: NERA analysis of uSwitch and Ofgem data. A simplification in Ofgem’s model may also introduce a systematic bias from the level of consumption. NERA’s model used for our own calculations of margins allowed consumption to vary across regions, suppliers, single-fuel and dual-fuel status, and tariff types. If customers optimise such that on average they choose cheaper tariffs for their level of consumption, it follows that allowing consumption to vary between tariffs and suppliers would result in lower bills overall. In other words, Ofgem models the bill for an average customer rather than the average bill and ignores the ability of customers of all types to make savings by selecting a lower bill for their consumption level, which will tend to bias its revenue estimates upwards. "Typical" Consumption SMI Consumption Difference £/year £/year £/year Cheapest Variable 1,137 1,268 131 Cheapest Fixed 923 1,025 102Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 11 A.2.2. “Blended” tariffs may place too much weight on more expensive deals Ofgem calculates the price at which large suppliers sell energy using a “blended” tariff, based on a mixture of the fixed and variable tariffs that are available to consumers.28 It uses quarterly statistics, reported by DECC, which suggest that only 64% of electricity customers and 61% of gas customers are on standard tariffs. The remainder are on either (1) online tariffs or (2) fixed tariffs. Ofgem’s published methodology implies that, in constructing its “blended” tariff, it places equal weight on all non-standard tariffs offered to customers.29 Since energy customers that switch tariff are driven primarily by price, lower tariffs will tend to attract more customers and hence lead to lower levels of revenue per customer. As a result, Ofgem’s method will tend to overstate the revenues that suppliers receive by placing too much weight on more expensive fixed price tariffs that fewer customers have signed up to. To quantify the likely effect this bias may introduce, we examined the difference between the cheapest and most expensive 12 month fixed price tariff offered by one of the large energy suppliers for an example address and for the consumption levels assumed by Ofgem.30 We found that the cheapest fixed electricity tariffs offered by the large suppliers varied between £466 and £585 per annum, while gas tariffs varied between £554 and £780. In Table A.2 we use an example to demonstrate how accounting for the range of fixed tariffs offered by suppliers, without accounting for the amount of energy sold through each tariff, will tend to overstate revenue per customer in the SMI. We take the unweighted average of six electricity tariffs (i.e. we assume a market share of 16.6% for each) to give a simple average of £522 per account, and do the same for gas (assuming a 20% market share for each) to give a simple average of £644 per account. We then construct some example weightings that might result from customers’ tendency to choose the cheapest deal, placing a weight of 30% on the cheapest and as little as 5% on the most expensive. With these example weightings the cost per customer account falls by £20 to £502 for electricity and £32 to £612 for gas. Finally, we take the proportion of customers on fixed tariffs used by Ofgem in its “blended” tariff to translate this into an overall effect on the revenue reported in the SMI. In the case of this particular illustrative example, Ofgem’s method results in an overstatement of average revenue of £5 per electricity customer and £8 per gas customer. 28 Ofgem (2014), para. 2.4, page 4. 29 “We weight the average tariff according to the split between the numbers of customers on a standard tariff and all other tariffs in the market”. Ofgem (2014), para 3.9, page 9. 30 Data retrieved for a London postcode from uSwitch.com on 26 January 2015.Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 12 Table A.2 Placing Too Much Weight on More Expensive Tariffs May Overstate Revenue Source: NERA analysis of uSwitch data. “Fixed proportions” are from DECC, Quarterly Energy Prices, Table 2.4.2 and 2.5.2, as used by Ofgem. A.3. Wholesale Costs Ofgem bases it analysis of the wholesale cost on an assumed 18 month-ahead hedging strategy. For example, gas and power to be delivered from 1 January 2015 to 31 December 2015 is procured via contracts for future delivery, with equal increments bought in each of the previous 18 months (December 2014, November 2014 etc.) A.3.1. Hedging strategies from 2008 may no longer be applicable Ofgem bases its choice of an 18 month-ahead hedging strategy on work carried out for the Energy Supply Probe in 2008.31 Even at the time, Ofgem only argued that an 18 month hedging strategy was representative of the market as a whole. In practice, suppliers pursued a range of different hedging strategies over longer and shorter time periods. It is not clear that an 18 month strategy would accurately measure industry wholesale costs in 2008 or now. In the intervening period of nearly seven years, the pricing policies and need for hedging risks according to a particular profile is likely to have changed. At the time of the Probe the practice of offering consumers fixed tariffs was less widespread than it is today. Ofgem estimated that roughly one in seven households were on a fixed tariff of some form (roughly 15%) at the time of the Probe.32 As of September 2014, figures produced by DECC suggest that 26% of gas customers and 24% of electricity customers are now on fixed or capped tariffs and 36% of electricity and 39% of gas customers are now on non-standard tariffs.33 Fixed tariffs may require longer-term hedging strategies to provide greater stability in fuel 31 “Based on our work for the Probe, we feel [18 months] is the most appropriate hedging strategy to use.” Ofgem (2014), page 12. 32 Ofgem (2008), Energy Supply Probe – Initial Findings Report, 6 October 2008, para 1.15 page 8; paras. 3.30-3.31, page 41. 33 DECC, Quarterly Energy Prices, September 2014, Tables 2.4.2 and 2.5.2. Tariff Notes Electricity Gas E.ON £466 - Fixed 1 Year (No gas tariff in this area) nPower £490 £565 Fixed Energy Online April 2016 S P £496 £554 Online Fixed Price Energy February 2016 EDF £521 £580 Blue + Price Promise May 2016 British Gas £572 £780 Fix & Fall June 2016 SSE £585 £739 Fixed Price March 2017 Simple Average £522 £644 Equal weightings Weighted Average £502 £612 Example unequal weightings (5-30%) Difference -20 -31 Simple minus Weighted Average Fixed Proportion 24% 26% Proportion Used in "Blended" Tariff Effect on SMI -£5 -£8 Change in SMI RevenueEnergy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 13 costs and/or strategies which fix fuel costs at the date of contract signature. In principle, other market developments might point to shorter term hedging. In the current climate of falling wholesale costs, suppliers with longer term hedging strategies or who had already hedged fixed price contracts would face higher wholesale costs for longer than Ofgem assumes, reducing net margins relative to those indicated by the SMI. A.3.2. Longer hedging strategies would currently exhibit higher costs We recreated the method for calculating wholesale costs described by Ofgem in its SMI Methodology. 34 From seasonal gas and electricity forwards, and Ofgem’s assumption about the annual profile of domestic demand, we calculated the weighted average price in any given month of supplying a customer for a subsequent 12 months (i.e. PSummer x % DemandSummer + PWinter + % DemandWinter). Following Ofgem, we assumed that 70% of domestic electricity consumption can be met by baseload contracts, while 30% must be met by peakload. We therefore calculated a measure of the “unhedged” price of supplying a new customer for the next twelve months. In practice, suppliers tend to purchase the gas and electricity they need to supply their customers “rateably” i.e. several months or years in advance of delivery. We therefore also calculated a “hedged” price, using the 18 month-ahead purchasing period assumed by Ofgem in the SMI. Figure A.1 illustrates the recent trends for an 18 month hedging strategy for gas and power respectively. While gas wholesale prices have fallen rapidly over the last twelve months, by 26%, suppliers purchasing gas under an 18 month-ahead hedging strategy have only seen their wholesale costs fall by 10%. 34 Ofgem (2014), Appendix 2. We use seasonal price assessment data available from Platts, a different provider to that used by Ofgem. Therefore, there are small differences in our calculations.Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 14 Figure A.1 Unhedged Gas Prices Have Fallen by 26% since January 2014, While Hedged Prices Have Only Fallen 10 % Source: NERA analysis of Platts data. Figure A.2 provides the same analysis for power (a 70:30 mixture of baseload and peakload). Wholesale costs have fallen much less, with hedged prices only 11% lower than in January 2014, as a result of which hedged prices have fallen by only 1%. Figure A.2 Unhedged Power Prices Have Fallen by 11% since January 2014, but Hedged Prices Have Only Fallen 1% Source: NERA analysis of Platts data. The increasing proportion of fixed price customers among the large suppliers suggests that a range of hedging strategies, including purchasing further in advance of delivery, may be 0 10 20 30 40 50 60 70 80 £/MWh Hedged Unhedged 0 10 20 30 40 50 60 70 80 p/therm Hedged UnhedgedEnergy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 15 appropriate. In Table A.3 we show the effect on a pounds per customer basis of hedging strategies of different lengths. The difference between a 12-month and 24-month strategy is £19 per gas customer and £8 per electricity customer. Suppliers who had hedged according to an 18-month contract would have additional wholesale costs of £7 per gas and £3 per electricity customer as compared to a 24-month strategy. Table A.3 The Cost of Hedging Gas and Power Purchases Longer in Advance is £8 -19 per Customer Source: NERA analysis of Platts data. A.3.3. Ofgem’s method does not take account of the cost of “shape” Ofgem’s SMI methodology assumes that energy suppliers buy all of their electricity and gas ahead of time in seasonal and quarterly blocks respectively. In reality, suppliers face further costs as demand varies within seasons and quarters, which require them to buy and sell electricity and gas to take account of the time-varying pattern of demand. In our previous reports for Energy UK, we estimated the cost of shape with reference to the contracts that suppliers must purchase to meet customer demand and the prices that suppliers must pay for these contracts. Our calculations took account of information from public data sources, our own estimates, and the response to our data requests to suppliers. We provide full details on the calculation method in our previous work.35 As far as we can tell from the published documentation, Ofgem’s work still excludes the costs of profiling wholesale gas and electricity to customer demand, whereas the cost of shaping wholesale electricity and gas to meet demand can be substantial. In our last report 35 NERA (2009), Energy Supply Margins At Current Prices, 7 December 2009, page 13. Hedging Strategy Gas Electricity Months £ per Customer £ per Customer 12 £297 £204 13 £299 £204 14 £301 £205 15 £303 £206 16 £305 £207 17 £307 £208 18 £309 £209 19 £310 £210 20 £312 £211 21 £313 £212 22 £314 £212 23 £315 £212 24 £316 £212 Range £19 £8Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 16 for Energy UK on supply margins in Great Britain, we estimated the cost of shape to be around £22 for electricity-only, £5 for gas-only and £23 for dual fuel customers.36 A.4. Operating Costs Suppliers’ costs also depend on operating costs which do not relate directly to the purchase and sale of gas and electricity (“indirect” costs). These include “customer service, staffing, IT, sales and marketing, billing, metering costs (including traditional and smart metering) and bad debt costs”.37 For the purposes of the SMI, Ofgem uses data from the most recent CSS, which in this case covers the calendar year of 2013, and adjusts it for inflation as measured by RPI. In doing so, Ofgem assumes that operating costs are roughly flat in real terms. There is a historical basis for this assumption, as indirect operating costs disclosed in the CSS have slightly decreased relative to RPI inflation. In practice, supplier’s costs may grow or shrink in nominal terms over the next twelve months. A.4.1. Fixed costs per customer may rise as large suppliers’ market shares fall Historical data is not likely to provide a robust basis for forecasting future costs because of recent trends in the domestic supply market. From January 2011 to October 2013, the combined market share of the large suppliers gradually decreased from 99.4% to 97.2%. Since then, it has fallen dramatically as many customers have switched to smaller suppliers. The most recent available data from July 2014 gives large suppliers a combined market share of 92.7%.38 Whilst in the long run suppliers will be able to scale their operations up or down with the size of their customer base, they are unlikely to be able to do so in the short run. Given that many operating costs are fixed in the short run, falling customer numbers will result in higher costs per customer. Ofgem estimates large suppliers’ indirect operating costs to be £79 per customer in the 2014 calendar year. If 100% of indirect operating costs were fixed (an upper bound assumption), the recent fall in the market share of large suppliers since the release of the 2013 CSS suggests that indirect costs per customer would have risen to £82 per customer (before accounting for inflation). A.4.2. Indirect costs may not increase in line with inflation Furthermore, there is little reason to believe that changes in many of the components of indirect operating costs are driven by general inflation. For example, the cost of bad debt should only be correlated with inflation over a long period of time. The total cost of a supplier’s bad debt is the cost per customer repaying a debt or in arrears multiplied by the number of customers in those states. The cost per customer is driven by the average bill per customer, which in turn is correlated with inflation over a long period of time. 36 NERA (2014), Energy Supply Margins :Update January 2012, 23 January 2012, page 4. 37 Ofgem (2014), Supply Market Indicator – Methodology, 27 November 2014, page 15, para. 3.21. 38 Cornwall Energy (2014), Competition in British household energy supply markets: an independent assessment, October 2014, pages 13, 15, 18, 20.Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 17 However, the number of customers repaying a debt (or not, as the case may be) may bear little relationship to the level of inflation. Rather, this is more likely to be driven by other macroeconomic trends, such as unemployment, income inequality and others. Another key component of indirect operating costs comes from customers switching suppliers. This is also unlikely to be strongly correlated with inflation, and will instead depend on the switching rate (or the “churn rate”). The end of 2013 and the beginning of 2014 saw a sharp increase in the churn rate, resulting in higher costs for suppliers.39 The churn rate has decreased since then, but the falling market share of large suppliers suggests that the churn rate may be more volatile than in the past and could increase substantially in the future. A.5. Network Costs and Social Obligations A.5.1. Network costs Ofgem’s Methodology sets out its calculation of distribution (DSUoS) and transmission (TNUoS) charges for electricity and gas, as well as the balancing (BSUoS) charges for electricity only. In each case, Ofgem’s relies on its assumed level of consumption and works out the capacity and commodity charges (or standing and unit charges) for delivering electricity and gas to a local distribution network.40 In the case of BSUoS, Ofgem applies the most recent £/MWh charges and applies them to domestic electricity consumption. This method is subject to a number of uncertainties that are likely to reduce the accuracy of the per customer figure Ofgem derives. Electricity Consumption: as noted in Section A.1, Ofgem’s calculation ignores the 5 million Economy 7 electricity meters in Great Britain, via which customers generally consume 6,000kWh per annum (rather than 3,800kWh assumed for the average customer on an unrestricted meter). Out of a total of 27 million meters, this suggests that Ofgem is excluding 11% of annual domestic electricity consumption and correspondingly understating the higher level of transmission charges that suppliers face to serve their entire pool of customers. Weighting by Region: Ofgem’s Methodology implies (but does not state) that it uses an unweighted average of transmission and distribution charges by network region or local distribution zone.41 Unweighted distribution charges will not reflect the average costs faced by suppliers in serving domestic customers, up or down depending on the correlation between higher tariffs and customer numbers connected to each distribution network. If Ofgem is using an average by total customer numbers in each region then 39 Cornwall Energy (2014), Competition in British household energy supply markets: an independent assessment, October 2014, page 13, 15, 18, 20 40 Ofgem (2014), Supply Market Indicator - Methodology , page 14, paras. 3.17-3.18. 41 For example: “To calculate distribution charges we multiply the unit rate by annual consumption and add the annual standing charge to arrive at a total annual charge for each electricity network region. We then take an average of these figures to use as the electricity distribution cost component within the SMI”. Ofgem (2014), Supply Market Indicator - Methodology , page 14, para. 3.18.Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 18 this is also potentially problematic. The SMI is a measure of profits of the large energy supply companies, who face varying degrees of competition from small suppliers in different regions. If Ofgem takes a weighted average using a weighting that does not reflect the customers of the large suppliers, then this will undermine the forecasts of costs that are produced. We have not attempted to quantify these biases in this report. However, they provide further illustration that the SMI exercise relies on highly uncertain and potentially biased methods to go from historical and factual measures of profitability to estimated and speculative forwardlooking estimates. A.5.2. Environmental and Social Obligations Four environmental and social obligations that suppliers pass through to customers are estimated by Ofgem. These are:42 Energy Company Obligation (ECO): the total annual average cost of providing support for energy efficiency installations in domestic premises, divided by the total number of gas and electricity customers; Renewable Obligation Certificates (ROCs): the cost of purchasing ROCs (at the “buy out price” in £/MWh) from qualifying generators to meet the total annual requirement in Great Britain, multiplied by a domestic customer’s proportion of total electricity demand; Feed in Tariffs (FiTs): the total annual cost of providing support to domestic customers with installations (e.g. solar panels) that qualify for FiTS, divided by the total number of electricity customers; and Warm Home Discount (WHD): the total annual average cost of providing discounts to qualifying homes, divided by the total number of gas and electricity customers. In the case of ROCS, FiTS and WHD, Ofgem’s method relies on historical costs (estimated or actual) and inflates them using RPI to give a forward-looking cost per customer. In the case of the ECO, Ofgem has selectively chosen a forecast that averages incurred costs (which have been low) with forecast future costs (which are due to rise) and potentially understating the cost of the obligation over the next twelve months. Ofgem uses an estimated of the annual cost of the ECO, between April 2014 and March 2016, of £787 million. DECC estimates that delivery costs will not be smooth over that period, stating that “we estimate that the average annual delivery costs to be passed through to consumer bills … [is] between £787 million and £820 million for the two year period to 31 March 2017”.43 DECC’s forecasts suggest delivery costs for the ECO are set to rise by £137 million between the 12 months beginning April 2014 and April 2015, which will not be reflected in the SMI. Ofgem’s choice of a smoothed estimate of the annual cost of the ECO also ignores the fact that the cost of the policy is likely to rise over time, as suppliers face increasing pressure to 42 Ofgem (2014), Supply Market Indicator - Methodology , page 16, para. 3.26. 43 DECC (2014), The Future of the Energy Company Obligation: Final Impact Assessment, 27 October 2014, page 30, para 121.Energy Supply Margins: Commentary on Ofgem's SMI Components of the SMI NERA Economic Consulting 19 meet their delivery targets before the delivery deadline. Data published by Ofgem shows that although the first phase of the ECO (from January 2013 to March 2015) is 81% complete, suppliers have only delivered 47% of their Carbon Saving Obligation (CSO) and 54% of their Carbon Saving Community (CSC) obligations (the two largest obligations under the three part policy).44 Clearly, delivery rates and hence costs of the obligation must increase if the targets are to be met on time and in full. Ofgem’s choice of a smoothed cost forecast is therefore likely to understate the costs to suppliers of meeting their obligations, although the exact magnitude of the bias is hard to quantify. 44 Ofgem E-serve (2015), Energy Companies Obligation – Compliance Update, January 2015, page 2, fig. 2.Energy Supply Margins: Commentary on Ofgem's SMI NERA Economic Consulting 20 Report qualifications/assumptions and limiting conditions This report is for the exclusive use of the NERA Economic Consulting client named herein. This report is not to be reproduced, quoted or distributed for any purpose without the prior written permission of NERA Economic Consulting. There are no third party beneficiaries with respect to this report, and NERA Economic Consulting does not accept any liability to any third party. Information furnished by others, upon which all or portions of this report are based, is believed to be reliable but has not been independently verified, unless otherwise expressly indicated. Public information and industry and statistical data are from sources we deem to be reliable; however, we make no representation as to the accuracy or completeness of such information. The findings contained in this report may contain predictions based on current data and historical trends. Any such predictions are subject to inherent risks and uncertainties. NERA Economic Consulting accepts no responsibility for actual results or future events. The opinions expressed in this report are valid only for the purpose stated herein and as of the date of this report. No obligation is assumed to revise this report to reflect changes, events or conditions, which occur subsequent to the date hereof. All decisions in connection with the implementation or use of advice or recommendations contained in this report are the sole responsibility of the client. This report does not represent investment advice nor does it provide an opinion regarding the fairness of any transaction to any and all parties.