Supplementary written evidence submitted by FTI Consulting (GRI0126)

1. On 7 February 2024, I served as a public witness to the Energy Security and Net Zero Committee (“ESNZ”) on the topic of “a flexible grid for the future”. My evidence was partly based on our independent assessment on behalf of Ofgem of the impact of locational wholesale electricity pricing in GB, published in October 2023.[1] In this long-standing engagement, which I led in my capacity as Head of EMEA Energy Transition at FTI Consulting, we found the following:

1. £28 billion to £51 billion of consumer benefits in GB between 2025 and 2040 in a nodal market design, and £15 billion to £41 billion in a zonal market design.
2. Consumers in each GB region benefitting, although the size of benefits differs across GB due to variation in wholesale electricity prices between different regions.
3. £13 billion to £24 billion of societal benefits in GB, which includes changes to generator revenues, in a nodal market design, and £6 billion to £15 billion in a zonal market design.
4. Improved utilisation of flexibility assets in dispatch, meaning lower spend on managing constraints and potential very significant savings in transmission investment.
5. 65 to 100 MtCO2 of reduced emissions in GB between 2025 and 2040 in a nodal market design, equivalent to £12 billion to £18 billion of additional benefits.
6. Significant changes to asset revenues leading to “winners and losers” among different stakeholders.
2. As a follow-up to that session, I have been asked to provide supplementary evidence on two specific topics. These are:

1. a further explanation of examples of other jurisdictions that have implemented a form of locational wholesale electricity pricing; and
2. in the GB context, how existing Contracts-for-Difference (“CfD”) subsidy schemes for renewables would be affected if locational wholesale electricity pricing was introduced, and the implications for consumers.
3. I discuss each of the two topics requested in turn.
4. I also note that one of the other witnesses at the Committee, Mr Gowdy, has submitted an extensive report to the Committee, since the hearing by way of a follow-up. I comment briefly on this, noting that I disagree with nearly all of the points made in Mr Gowdy’s submission.

A. Further explanation of the experience of other markets

5. One of the requests following my oral evidence to the Select Committee was to explain the experience of other locational pricing markets further.
6. Locational wholesale electricity pricing is a well-established feature of electricity markets since the wave of liberalisation in the 1980s and 1990s. Unlike GB and some European countries such as France and Germany, many other markets have adopted two broad approaches to locational pricing, often expressed as “zonal pricing” or “nodal pricing”.
7. At the outset, I highlight that locational pricing is becoming a more common feature in liberalised electricity markets worldwide. This is shown in Figure 1 below, which summarises the evolution of installed generation capacity in liberalised OECD markets, and the split of those markets between national, zonal and nodal market designs.

Figure 1: Evolution of installed capacity under different market design options across the OECD countries with liberalised electricity markets

8. We make the following observations:

• In 2000, about 68% of the installed capacity of liberalised electricity markets in the OECD operated under national pricing. These included the nascent liberalised markets of Europe, as well as some regions of the US. Zonal pricing existed in a small number of regions, such as Australia, Norway and Japan. Nodal markets were at the time relatively unusual – operating only in New Zealand as well as PJM and NYISO in the US.
• By 2020, most regions in the US, including California and Texas, had transitioned to nodal market designs as a result of perceived shortcomings in earlier approaches. Some national markets also transitioned to zonal markets, for example Sweden adopted a four-zone market in 2011.
• Apart from the expansion of NETA to include Scotland in 2005, and a reconfiguration of one zone in Australia in 2008, we are not aware of other examples where the locational granularity of wholesale electricity markets has reduced.
9. Table 1 below presents some relevant statistics from the GB electricity market, and selected electricity markets that have implemented some form of locational pricing around the world. Several of these countries have aspects directly comparable to GB.

Table 1: Relevant statistics from selected electricity markets

Note: Some jurisdictions, such as CAISO, Sweden and New Zealand, have significant renewable resources such as hydropower and geothermal not included in the two right-hand side columns.

10. As such, I provide a brief overview for selected jurisdictions from the above, explaining their rationale for transitioning to locational pricing, recent observable trends in the market, and key implications for the GB electricity market. I cover the following markets:

1. ERCOT, Texas
2. CAISO, California
3. Svenska Kraftnat (“SvK”), Sweden
4. New Zealand

a. ERCOT, Texas

11. The electricity market in Texas, often referred to as the Electric Reliability Council of Texas (“ERCOT”) (the system operator), is similar to the GB electricity market in some ways. We set out a few key comparisons:[12],[13]

• Although ERCOT only serves about 26m people, the size of the ERCOT and GB electricity markets are not too dissimilar with peak demand at approximately 80GW and 60GW respectively.
• Like GB, ERCOT has been decarbonising the power sector through significant wind investments, at c.37GW compared to c.30GW in GB as at end-2022. Most of these wind investments are located in remote areas, such as West Texas in the case of ERCOT, which is distant from several demand centres in the north, notably Houston.
• Both electricity markets also have relatively limited interconnection with neighbouring regions, in the case of ERCOT, to the rest of the US or Mexico.
12. However, the experiences of ERCOT and GB have diverged historically. In 2010, ERCOT transitioned to a nodal market from a zonal market (which was adopted in 2002). Following extensive stakeholder engagement and Cost Benefit Analyses, a nodal market was deemed to be necessary to resolve congestion, utilise generation and transmission resources more efficiently, and integrate increasing renewables, among other benefits.[14]
13. ERCOT has often been characterised as a very competitive energy market, encouraging competition in generation investments without the need for extensive policy interventions such as a capacity mechanism.[15] In more recent years, since transitioning into a nodal market in 2010, renewable investments have picked up significantly, notably wind but more recently solar and battery storage.
14. ERCOT also has a competitive retail market, with about 75% of consumers under a competitive tariff contract.[16] As a nodal market, demand side response (i.e. the ability to reduce or shift demand) has made a sizeable contribution to the operation of the system, participating in a similar way to generators to support energy resilience and security of supply.[17] A nodal market, which sets locational prices at 5-minute intervals, provides both the locational and temporal granularity for nimble technologies such as demand-side response to contribute more effectively to balancing the system.[18] I show the average annual wholesale price below across each zone (represented as the average nodal price in an area) in Figure 2 below.[19]
    

Figure 2: ERCOT average annual real-time energy market price by zone, USD/MWh

15. While some stakeholders argue that locational pricing reforms in GB would be detrimental to generation investments, and in turn, progress towards Net Zero, the experience in ERCOT suggests otherwise. We set out the annual installed capacity additions in ERCOT below in Figure 3, and Wind installed capacity by ERCOT zone as of end-2022 in Figure 4.[20] Figure 5 shows the location of wind turbines in Texas.[21]

Figure 3: Additional installed capacity per annum by generation technology in ERCOT, GW

Figure 4: Wind installed capacity by ERCOT zone as of end-2022, GW

Note: Figure 4 above shows wind generation capacity by ERCOT zones based on several sources. As the ERCOT grid does not cover the entire state, the Panhandle in Northwest Texas is mostly part of the Southwest Power Pool (“SPP”) grid, and so we use Potomac Economics: 2022 State of the Market Report to reflect installed capacity in the ‘North’.

Figure 5: Location of wind turbines in Texas by capacity

16. We make the following observations.

• The new capacity since 2010 has mostly been wind investments, making up 37GW of total capacity in ERCOT as at end-2022.[22] As shown in Figures 4 and 5 above, most of these wind investments are located in West Texas, a relatively remote area which is distant from several demand centres in the North, most notably Houston.
• A considerable proportion of ERCOT’s wind investments have also been made in North and South Texas, areas with less network congestion. In total, out of the 37GW of wind capacity as at end-2022, 8GW is located in South Texas and 3GW in North Texas.[23]
• More recently, significant solar and battery storage investments have also been made. As shown in Figure 3 above, from 2018 to 2023 there were investments of around 12GW of wind, 8GW of solar and 2GW of battery storage, representing about 30% of total installed capacity across all of Texas. In the same period, the GB electricity market, which is a similar size to ERCOT (in terms of energy consumption), invested in c.6GW of wind, 1GW of solar and 2GW of battery storage.[24],[25]
17. ERCOT is therefore marked as a useful case study to show that an alternative market design can achieve a successful integration of generation resources (across a range of types, particularly wind) and flexible demand. Like other jurisdictions, however, ERCOT, is not perfect, and it faces many challenges in its energy transition, for example in the need to bolster resilience to bad weather and to increase interconnection. However, these challenges are features of the context it operates in; they are not related to having a locational pricing market.[26]

b. CAISO, California

18. The electricity market in California, often referred to as the California Independent System Operator (“CAISO”) (as the system operator), transitioned from a zonal market, established in 1998, to a nodal market in 2009. CAISO’s peak load is over 50GW, slightly smaller than GB’s peak load of around 60GW.
19. While there are several insights from the history of CAISO, such as its energy crisis of 2000-2001 which was driven by poor market design leading to rampant market manipulation and ineffective treatment of congestion, the key recent lessons for GB relate to its successful roll-out of renewables – mostly solar and batteries.
20. In the past six years (2017 to 2023), 8GW of solar, 1.5GW of wind, 6.5GW of battery and 1GW of renewable-battery hybrid were installed. This is shown in Figure 6 below.[27]

Figure 6: Additions to CAISO market participation by fuel type

21. As shown in Figure 6 above, most of the new battery capacity investments have been made since 2021, to store energy from high solar generation when it is sunny to serve demand in other parts of the day, such as the evening.
22. More recently, electricity markets adjacent to CAISO have joined CAISO’s real-time nodal market, known as the Western Energy Imbalance Market (“WEIM”). This is shown in Figure 7 below.[28] The initiative has been widely regarded as a success, with estimated benefits of c.$1.7 billion in 2023.[29] Further expansion of the nodal CAISO market is anticipated in the forthcoming period. [30]

Figure 7: Active participants in the WEIM

c. Svenska Kraftnät, Sweden

23. In 2011, SvK, the Swedish transmission owner and system operator, divided the Swedish electricity market into four zones, as shown in Figure 8 below.[31] The boundaries of the zones were defined by where the main “bottlenecks” on the grid were located.

              Figure 8: Map of Swedish electricity market, and Nordic and Baltic bidding zones in 2022

24. In some ways, the Swedish energy landscape is analogous to GB. The north of Sweden is marked by rich renewable resources (with some heavy manufacturing industries), while most of the population is located in southern Sweden. Transmission capacity to connect northern and southern Sweden is limited, which risks shortfalls in generation in southern Sweden.
25. The reform followed a European Commission (“EC”) finding that SvK abused its dominant position by “curtailing capacity on the Swedish interconnectors when it anticipated internal congestion within the Swedish transmission system”, particularly to Denmark.[32] This arose in periods when southern Sweden had excess demand, but limited transmission capacity to convey generation from northern Sweden. In some of these periods, due to a low single uniform price across the country, interconnectors from south Sweden were scheduled to export which would exacerbate congestion within Sweden, and in particular north-south flows of electricity.[33]
26. After the EC’s judgment, Sweden decided to implement four zones where each zone would have its own locational wholesale price. The southern zone, which often has excess demand, typically has a higher locational wholesale price, improving how the interconnectors would be scheduled (i.e. importing more often to Sweden when there is congestion between north and south Sweden).
27. In GB, because the wholesale market does not account for transmission constraints when scheduling supply to meet demand, interconnectors to France are often scheduled to export from GB to France when, the physical reality of the system is such that they should instead be importing to meet excess demand in south GB. In our modelling for Ofgem, we found that interconnector flows were expected to be scheduled the “wrong way” up to 30% time in 2030.[34] In these hours, the GB energy system is essentially scheduled as a highway, paid by consumers, to convey imported power from Norway and renewables generation in Scotland to France – but frequently without the necessary transmission infrastructure to actually deliver these flows.[35]
28. In Sweden, following the wholesale market reform, prices between the four zones have increasingly diverged, particularly since 2020. In 2022, the average annual zonal price in the northernmost zone in Sweden was 0.63 SEK/kWh, but 1.62 SEK/kWh in the southernmost zone.[36] Figure 9 shows weekly prices across the Swedish zones in 2022.[37]

Figure 9: Average annual wholesale prices in Sweden in 2022 and 2023 by price zone, SEK/kWh

29. Additionally, in recent years, SvK has received significant congestion rents (c.SEK 49 billion in 2022) which arises due to price differentials across two zones (revenues collected exceed payments to generators across two zones where there is insufficient transmission capacity and a price differential). This congestion rent is typically used to reduce transmission charges, and so constitutes an additional benefit to Swedish consumers of the move to locational pricing.
30. There is also evidence that locational wholesale prices in Sweden have affected the location decisions of new generation and energy-intensive industrial consumers. For example:

• A peer-reviewed academic study has found that wind investors have responded to locational price signals in Sweden, with greater investments being made in higher price zones.[38]
• As I noted in the Committee, in September 2023, H2 Green Steel announced it would be building the world’s first large-scale green steel plant and Europe’s first giga-scale green electrolyser in Boden Sweden. H2 Green Steel states “we chose to locate our production in Boden and Norra Svartbyn thanks to the ample access to renewable energy at competitive prices in the long term, proximity to high-quality iron ore, vital infrastructure such as the Malmbanan railway and the Port of Luleå, and leading expertise in the region – an ideal environment for our venture [emphasis added]”.[39] This supports the comments I discussed with Rt Hon Mark Garnier in the Parliamentary session, that while electricity prices may be one of several factors affecting location decisions, there is inevitably a trade-off between low electricity prices in more remote areas versus the availability of other factors such as specialised labour or transport infrastructure in densely populated areas.[40]

d. New Zealand

31. The electricity market in New Zealand is considered to be the first nodal market, beginning trading in October 1996.
32. The electricity market in New Zealand has some similarities to GB with rich renewable resources in one part of the country (the South Island) but with most of demand in another (the North Island).
33. Hydropower makes up over half of New Zealand’s electricity generation, and is generally plentiful from Spring onwards as melting snow fills the reservoirs. This is correlated with lower demand, and can lead to excess generation in summer and excess demand in winter.
34. Alongside this, New Zealand’s transmission grid faces some particular issues; it consists of a long “backbone” along the length of the two islands, with few alternative paths between generators and demand. There is also significant distance between generation and demand, which can lead to large electrical losses. Finally, the North and South Islands have limited interconnection. In this context, locational pricing was seen as essential to manage potentially very differing supply and demand conditions that arise simultaneously at different locations across the system and to manage efficiently flows across the New Zealand transmission network.
35. Wind and solar penetration has increased in recent years in New Zealand; in 2010, wind and solar made up around 6% of installed capacity, and by 2022 it was around 14%.[41] Geothermal, wind and solar generation have made up the bulk of the transition away from coal generation in New Zealand.
36. Figures 10 and 11 below show a map of each region in New Zealand and the average nodal electricity price in these regions in 2022.[42] Consistent with theory, prices on the Lower South Island, where low-marginal cost hydro generation is more plentiful, are lower; prices in the North, particularly the Upper North Island, are generally higher, as the electricity is often supplied by more expensive gas generation.

Figure 10: Map of New Zealand Regions

Figure 11: Average nodal prices in each region of New Zealand in 2022, NZD/MWh

B. Existing generators with CfDs in a locational market

37. I have also been requested to provide a note about the impact of locational prices on existing wind generation investments backed by renewable support schemes, and the resulting implications for consumers.[43] As most concerns relate to projects backed by the so-called CfD regime, I focus on the impact to wind generators under these CfD contracts.[44]
38. The CfD regime is currently GB’s primary mechanism to support new renewable investment, particularly offshore wind investments.[45] It works by providing the CfD contract holder with a fixed guaranteed electricity price (known as the “strike price”) over 15 years.[46] This means that if the prevailing market wholesale electricity price (known as the “reference price”) is lower than the strike price, wind generators receive a support payment to top up the price to the level of the strike price. Likewise, when the reference price is above the strike price, wind generators pay the difference between these two prices.[47] Consumers pay for any top-ups, and receive payments from generators when the reference price is above the strike price.
39. CfD generators are guaranteed the strike price for their output, but do not receive it when they are not generating. However, these generators have firm transmission access under the current market design – which means they are entitled to be paid constrained-off payments when they are scheduled to run but are instructed by the ESO to be curtailed due to transmission congestion. These payments are ultimately borne by consumers.[48]
40. In our assessment of GB transitioning to a locational market, we have assumed throughout that all existing CfD contracts and planned CfD contracts would be maintained and honoured (often termed as “grandfathering”).[49] This means that CfD generators would continue receiving the prevailing strike price over the term of the CfD contract irrespective of the locational wholesale price.
41. As most of the existing and planned CfD wind generators are located in Scotland and northern England, locational pricing will likely, in aggregate, increase the cost of funding these CfD schemes to maintain the prevailing strike prices. This is because wholesale electricity prices in northern GB are likely to be lower due to limited transmission capacity to export excess renewable generation in certain periods, thereby increasing the support payment to maintain the prevailing CfD strike prices. As such, in a locational market, CfD investors would be no worse-off in periods of generation, and by extension, consumers would also be no better-off in these periods.[50] Overall, this reduces the benefits to consumers of transitioning to a locational pricing market despite lower wholesale electricity prices in northern GB.
42. The impact to consumers and generators from maintaining existing and planned CfD contracts can be shown in Figure 12 below:[51],[52]

Figure 12: Cost-benefit assessment for a nodal market design relative to a national market design (2025-2040) – LtW (NOA7)

43. As shown in Figure 12 above, maintaining existing and planned CfDs would constitute a cost to consumers. In one of our scenarios, we estimate a £12 billion cost to consumers over the modelling period, which negates some of the consumer benefits (though these remain highly positive, at £51 billion). Conversely, however, producers would benefit through the CfD prices – the £12 billion represents a transfer from consumers to generators. Overall, this reflects the assumption that all existing and planned CfDs are grandfathered in a transition to a locational pricing market.
44. Overall, we would expect the CfD regime to provide very significant confidence to investors in revenues from low-carbon generation. However, I highlight here some of the changes generators might face in a locational pricing market:

• First, all generators, including CfD generators, would not receive compensation revenue for periods when they are instructed to curtail due to insufficient transmission capacity to export excess generation. This leads to a decrease in generator profits but a benefit to consumers and an improvement in overall economic outcomes.
• Second, once the 15-year term of the CfD contracts ends, it is assumed that these generators would only receive the locational wholesale electricity price where they are sited. For northern generators, this may be markedly less in a locational market than the status quo national market.
• Third, new CfD contracts (beyond the current planned auctions at the time of the assessment) are expected to be awarded to prospective generators at sites more advantageous for the overall GB grid where they may be curtailed less frequently (subject to various limits such as seabed leases).[53] For example, under locational pricing, we would expect that more future CfD contracts would be allocated to offshore wind projects in the Celtic Sea instead in the North Sea.
45. While these are our assumptions on how a locational pricing market may impact CfD generators, policymakers may choose to implement additional transitional and mitigation measures. For example, policymakers may choose to continue providing CfD generators some compensation for curtailment, perhaps on a temporary basis, or to extend CfD contracts beyond 15-years to avoid early closure. I consider such policies to be treated as equal and opposite transfers; that is for every additional £1 awarded to generators, consumer benefits reduce by £1.[54]
46. Overall, while maintaining existing and planned CfD contracts might reduce the overall benefits of locational pricing, expected consumer benefits are still very significant at £51 billion, as are expected socioeconomic benefits at £24 billion (as shown in Figure 12). This reflects a wide range of benefits produced in a locational pricing market that far exceed the cost, for example through the efficient operations of flexibility assets (e.g. interconnectors and batteries). There are also additional benefits we did not model, such as the potential to avoid a potentially material proportion of the costs of incremental investment in the transmission network. This benefit would be over and above the benefits the £51bn of consumer benefits cited above.

C. Response to Mr Gowdy’s supplementary evidence to the ESNZ Select Committee

47. Since the public session on 7 February, I note that one of the other witnesses at the Committee, Mr Gowdy, has submitted supplementary evidence to the Committee. I provide a short response in this section.
48. Mr Gowdy’s response is some 32 pages in length. I disagree with the majority of points he raises in his note, but I will spare the committee a full breakdown and highlight seven key points Mr Gowdy makes that are either factually incorrect, unevidenced assertions, or are presented in a distortive manner in favour of maintenance of the status quo.
49. Point 1: Role of FTI Consulting. In Section 1.3 Mr Gowdy states that “studies commissioned by the ESO and Ofgem have used the same consultants (FTI) who have essentially used the same model using the same scenario assumptions”. This is incorrect. Our work for ESO was one of four phases of work commissioned by the ESO to assist it in its consideration of market reform as GB transitions to Net Zero. FTI Consulting were selected to assist in the third phase of ESO’s work, and our outputs were shared with the industry in a series of workshops.[55] In phases 1, 2 and 4 of its work, ESO used consultants KPMG, Frontier Economics/LCP and Baringa respectively.[56] Furthermore our work for ESO was not quantitative in nature – there is no “same” model used in our work for ESO that has been published to date.[57]
50. Point 2: Choice of scenarios used in modelling. Many of those opposed to locational pricing have sought to undermine the credibility of our modelling work for Ofgem by arguing that we should have assumed different scenarios of the transition pathway to Net Zero from those that we used in our work. Indeed, Mr Gowdy spends many pages on the topic in his submission. I have three simple points in response:

• First, the scenarios selected were the most up to date FES scenarios at the time our work was undertaken. These FES scenarios are used by the ESO to plan the electricity system – including transmission network build decisions[58] – and have themselves been through an exhaustive consultation process.[59] These scenarios are also used by Ofgem to take regulatory decisions – for example the recent assessment of the benefits of more interconnectors used the FES scenarios in Ofgem’s evaluation process.[60]
• Second, the scenarios we used were selected by Ofgem after very extensive consultation with the industry through workshops. They were considered to represent a fair range of potential trajectories to Net Zero (with one focusing on the roll-out of renewables generation – particularly in Scotland, one with very significant increase in transmission, and one being more focussed on the roll-out of nuclear generation in the south of the country). Indeed, Mr Gowdy himself played a full and vocal role in this process at the time the work was undertaken.
• Third, we have been, and remain, entirely “scenario agnostic”. We have simply modelled the benefits of moving to locational pricing under the scenarios as agreed by our client, Ofgem. Indeed, as I have frequently stated, we remain happy to model any scenario that we are commissioned to do so by any client.
51. In his report, Mr Gowdy suggested we should have modelled a scenario that incorporates the so-called “System Transformation” scenario and the “Holistic Network Design” network upgrade. Mr Gowdy’s argues that, if we had modelled this scenario, we would see a very significant drop in the assessed benefits of locational pricing.[61]
52. I agree with Mr Gowdy on this point. The reason this would show reduced benefits of a move to locational pricing is simply because it conflates two Net Zero scenarios: 1) an assumed roll-out of nuclear generation in the south of the country; and 2) a very large expansion of the electricity transmission system to facilitate the flow of a large volume of renewables generation from Scotland to the south.[62]
53. In some ways, this argument of using a “double Net Zero scenario” goes to the heart of the issue. It is probably correct that if we, as a country, expend resources on two Net Zero scenarios rather than one, then the incremental benefits of locational pricing would be low. However, given even one Net Zero scenario is likely to cost the country tens, if not hundreds, billions of pounds, doing two of them in lieu of reforming the electricity market design – which is essentially no more than changing some words on pages and changing some computer systems which in turn alters the financial flows around the system that impacts participants’ production and consumption behaviours in a way that is beneficial to the system – is quite clearly an absurd proposition.
54. Point 3: Mistaken conflation of real time spot markets and forwards markets. A repeated error of Mr Gowdy in his response[63] and in the Select Committee itself[64] is his conflation of the real time spot markets and forward markets. The debate on market design is a debate on how real time electricity prices should be set, where the real time price represents the price at which sufficient supply of electricity is forthcoming from generators to meet the demand for electricity in a specified time period (of 30 minutes in GB). In the GB market, since 2001 the real time price for energy is set on a national basis. The day ahead price is often used as a proxy for this price.[65] The previous GB market design, known as the Pool which covered England & Wales and went live in 1991, also operated on a single national price.
55. By contrast, Power Purchase Agreements (“PPAs”), referred to by Mr Gowdy, are forward contracts that are set with reference to delivery at a future point in time. They are typically entered into by participants to reduce the risk of exposure to the potentially volatile real time price. Forward contracts exist in all market designs – be it locational or national – and also existed under the previous Pool market design in England & Wales. Hence, the conflation of the forwards and real time markets is simply a distraction and it is unclear to me why Mr Gowdy repeatedly raises this point.
56. Point 4: Mr Gowdy’s reference to the £12 billion figure in our work for Ofgem. Mr Gowdy attempts to argue that I have said that wholesale costs in aggregate will go up in any potential move to locational pricing. Mr Gowdy does this by highlighting one of the five impacts on consumers (and six on society as a whole) that we evaluated in our study,[66] the breakdown of which is shown in Figure 13 below.[67]

Figure 13: Overall Cost Benefit Assessment for a nodal market design relative to a national market design (2025-2040) – LtW (NOA7)

57. As shown in Figure 13 above, the chart itself makes very clear the £12.7 billion figure (the light blue column) cited by Mr Gowdy is more than offset by the £48.8 billion reduction in payment by consumers in constraint management costs (the red column). Essentially the £12.7 billion cost arises as some of the costs that are saved in the red column (specifically constrained-on payments to southern based generators) still need to be incurred under a locational market design, but now manifest themselves in the light blue column.[68] The difference between the height of these two columns therefore represents one of the potential benefits of a shift to locational pricing.
58. A further £27.1 billion is earned by congestion rents (the green column) which would normally accrue to consumers.[69] The key point to note is that all of the costs must be considered in their totality to conduct a comprehensive and meaningful Cost Benefit Assessment. Hence to cite one figure in isolation is highly misleading.
59. Point 5: Mr Gowdy’s repeated references to other LMP markets – notably Texas – are entirely incorrect. I am very fortunate to lead a global team of energy market economists and work extensively both in GB-style non-locational markets as well as locational markets. For example, in recent years I have led work and filed expert reports in the LMP markets of New Zealand[70] and Texas[71], led a large team assisting the regulators on market design issues in the zonal market of Australia,[72] and supported colleagues in FTI Consulting’s work in the Mid-West of the US. As such, I am familiar with the workings of such markets as well as our own market here in GB.
60. As I have discussed earlier in this paper, Mr Gowdy appears to have a very distorted view of LMP markets such as Texas. Statements such as “many developers […] could decide to delay their projects until they have revenue security and a guarantee that network capacity will be available. [That] is what has happened in other LMP markets, such as Texas”[73] are simply untrue and not borne out by the facts. For example, as I noted earlier in this paper, Teas has had a very significant roll out of generation in a variety of different locations since 2022. Indeed, Texas increased its wind capacity in 2023 by 4GW alone - 40% more than Britain over the same period.
61. Point 6: Mr Gowdy presents a very distorted view of the cost to customers for constrained-off payments. A frequent assertion by Mr Gowdy is that the “largest portion of constraint costs is not the payments to generators to turn down, but is in fact the payments to generators […] to turn up to replace the constrained generation”.[74] I note that Mr Gowdy also made a similar point in his presentation to the Select Committee on 17th January when he states that “Around 70-80% of constraint costs are to Turn UP energy!”. [75]
62. In my view, this constitutes a very serious and significant misrepresentation of the costs to GB consumers of constraint costs and also of the amount of money paid to constrained-off generators (which are typically wind generators in the north of the country).
63. Mr Gowdy is indeed correct to note that a large portion of constraint costs incurred by the ESO are to turn up generators – typically gas fired plant in the south of the country. The smaller portion is those costs paid to renewables generators as compensation for being curtailed. These transactions are undertaken by the ESO in the balancing mechanism – with the costs being recovered from consumers.[76]
64. However, in all of his citations, Mr Gowdy at no point refers to the fact that in nearly all cases, renewables generators that are constrained-off have also been paid in the wholesale market. Hence, constrained-off generators are the recipient of two payment flows from consumers – one payment in the balancing mechanism (paid by ESO and recovered from consumers) and one payment in the wholesale market (paid by retailers and ultimately recovered from consumers). But Mr Gowdy only refers to one payment flow to constrained-off generators. Figure 14 below provides a simple schematic diagram of the payment flows associated with constraining on and off wind generators under the current market design.

Figure 14: Payment flows associated with constraint costs under current market design

65. The key point to note is that, under typical market conditions, the largest proportion of money paid by consumers to constrained-off wind generators is not through the payments by the ESO but rather through the wholesale market.[77]
66. Because of the way the current market works – with the split between the non-locational wholesale market and the locational balancing mechanism, combined with the opaque nature of the contracts market - it is very difficult to assess the revenue constrained-off generators earn for not actually generating. However, through our modelling for Ofgem, we have estimated payments to wind generators sited in Scotland that they would earn over the period 2025 to 2040.
67. Our, very conservative, estimate is that Scottish wind generators would receive payments from consumers of £11 billion in the period 2025 to 2040 for not actually generating.[78] This comes in two forms – payments in the balancing mechanism of £1 billion (the payments referred to by Mr Gowdy) and payments in the wholesale market of £10 billion (the payments ignored by Mr Gowdy).
68. Given the large quantum of customer money that would, under the current regime, be paid to wind generators in Scotland for, quite literally, nothing, a full understanding of the financial implications of the current regime therefore needs to take account of both types of payment. I would discount entirely the view put forward by Mr Gowdy that the vast majority of costs to customers to resolve constraint costs are incurred by plant to “turn up” which, in my view, is a very serious mischaracterisation of the true impact on consumers of the current market design.
69. Point 7: Mr Gowdy’s list of “[o]pportunities to reduce the occurrence and cost of constraints”[79] represents a compendium of “Heath Robinson” style sticking plasters on the GB electricity market. In his submission to the Select Committee, Mr Gowdy provides “good news” in the form of a long list of reforms that might be adopted in lieu of locational pricing. I will resist the temptation to explain why each one of them will either be hugely costly to consumers – simply noting that they would retain the current approach of wealth transfers from GB consumers to constrained-off generators – ineffective, or both.
70. Instead, I will simply requote my statement at the Hearing that “[f]irstly some of those proposals are entirely untested relative to locational pricing, which has been in operation in many parts of the world for over 20 years. Secondly, under that regime, the constraint payments that generators in the north of the country continue to receive, which is money literally for nothing, would continue.”[80]

April 2024