Briefing paper
Review of Energy Policy 2023
Dec 2023
Authors: Simon Gill, Keith Bell and Callum MacIver
The introduction of Locational Marginal Pricing (LMP) is being considered as part of a broad debate about electricity market reform in Great Britain. This extended blog argues that whilst there may be theoretical benefits to LMP, there is much still to be thought about in terms of detailed design and how it could fit with other market arrangements.
There are huge challenges in delivering an LMP market in the near term that is well adapted to the GB system, so much so that it has the potential to put the UK Government’s commitment to decarbonise the electricity system by 2035 at significant risk. This finding has emerged from a major review of LMP that we have recently undertaken which includes an exploration of how LMP markets work in practice, the particular challenges of the GB context, and the interaction between the wholesale energy market and other elements of the market framework and wider electricity system.
In a previous blog we described the broad outline of an LMP market and its characteristics and laid out some of the arguments for and against. In the report that we’ve just published we dive into considerable detail looking at the components of an LMP based market and how they work together, discuss experience from existing LMP markets around the world and consider how that experience might relate to the GB context. Our findings are based on a review of published literature plus the views of experts from Britain, the US and Australia gathered through a set of ten semi-structured interviews. (An extended summary of those discussions is presented in a dedicated Stakeholder Insight report). Here we discuss some what we have explored and what we’ve found.
The UK Government’s Review of Electricity Market Arrangements (REMA) covers substantially more than the wholesale market for electrical energy and the potential introduction of LMP. In addition, it considers options for how best to support mass development of low carbon power and ‘flexibility’, options for reform of the Capacity Market, how ancillary services are procured, and how the system can remain operable. There are also major elements of reform needed that are beyond the scope of REMA such as how to speed up delivery of substantial upgrades to our transmission network – a challenge set for the Electricity Networks Commissioner appointed in July last year – and the need for reform of the retail electricity market alongside wholesale market reform.
The UK Government has recognised that many elements of electricity sector commercial and regulatory arrangements affect each other: the final impact of any option for reform of the wholesale electricity market will, to a very large extent, be dictated by its interaction with these other elements. For example, whilst many commentators highlight the risk that an LMP market could put on renewable generators it is important to remember that this could be at least partially offset by the re-design of Government-backed Contracts for Difference (CfDs) or alternative low carbon support mechanisms. The trouble is that this redesign is highly complex. LMP markets have never been combined with the type of low carbon support used in GB where the level of support varies inversely with a direct measure of the day-ahead wholesale price. In the US – from where most experience of the operation of LMP based markets is derived – renewables are typically supported by production tax credits which provide a flat rate, level subsidy per MWh generated on top of wholesale prices, similar to our legacy Renewables Obligation and Feed-in Tariff schemes. The number of combinations of choice of reference price, negative pricing rules, the interaction with Financial Transmission Rights (FTRs) and initial CfD auction design, lead to dozens of different design options each with its own set of incentives on market participants. (We illustrate five in our report to make the point that each can drive very different bidding strategies from renewables in a day ahead LMP market).
FTRs have been touted by proponents of LMP as a particular means of managing risk for market participants under an LMP market. However, our investigation reveals significant issues with existing implementations of FTRs in the US which question the validity of such claims. These include their short duration compared with asset lifecycles and potential for market manipulation. In addition, we identify the likely need for first-of-a-kind adaptations to FTR products to enable compatibility with variable renewable generation assets. Serious attempts to identify such practical issues and, more importantly, detail possible solutions have thus far been lacking from the GB debate.
Our conclusion is that, with significant structural interactions between an LMP wholesale market and other elements, simple assertions that LMP is definitely ‘good’ or definitely ‘bad’ should be avoided. The challenges of market reform are too difficult and too impactful to be treated as theoretical issues where practicalities and real-world realities can be ignored. Informed discussion is needed on the merits of different packages of reforms that include LMP compared with packages of reforms that don’t.
Debate around the increase in cost of capital that a move to LMP could create seems particularly heated, but good quality, publicly available analytical evidence is relatively limited. Assumptions being made by FTI Consulting in their ongoing work for Ofgem include relatively small cost of capital increases – 0.25 to 0.5 percentage points for renewables – with little visible justification of the values used. By contrast, many developers argue that the increase could easily be in the range of 2 to 3 percentage points, a level which would wipe out the large consumer and overall welfare benefits modelled by FTI. A recent report by Frontier Economics provides some supporting analysis which indicates potential cost of capital impacts in the same range. Work by the UK Energy Research Centre published in November 2021 also points towards potentially high values.
Whilst we understand that confidentiality and competition implications of sharing hard evidence on investment decisions make this a very challenging area to explore in public debate, it is critical that we don’t make a decision on major market reform without a significantly better understanding of the cost of capital implications than we have today. Ideally, the sector would move towards a shared consensus for the values used in modelling, or at least an agreed set of sensitivities and terminology to refer to them.
Drawing on our previous point, cost of capital is not just dependent on the wholesale market reform options chosen – LMP or otherwise – nor is it just dependent on wholesale reform and decisions for each element of our market framework in isolation. Rather, it is dependent on the overall combination of reforms including the critical design choices that will enable the wholesale market to interact with low carbon support, capacity adequacy, ancillary services and more. Reform packages that limit the impact of a switch to LMP on cost of capital may necessarily also mute the potential consumer welfare benefits. All stakeholders should acknowledge this point and engage with it.
There are two broad sets of outcomes that a move to LMPs is argued to create: impacts on siting of investments, and impacts on dispatch and utilisation of available resources.
The key part of the utilisation argument is, it seems to us, that decision making is centralised. The scheduling and dispatch of available resources – generation, storage, interconnector transfers and flexible demand – to meet demand can be most cost-effectively done through a centrally-run optimisation algorithm that reduces the inefficiencies arising from each market actor dispatching assets in their own portfolios independently of each other. However, a centralised, ‘gross pool’ also removes network users’ firm rights to access the network and would mark a departure from the existing ‘connect and manage’ regime for onshore generation. Access is granted only if the algorithm chooses you, and it only does that if its model of the system says that you can be accommodated as part of a least-cost day-ahead dispatch to meet demand. On a network with limited capacity, that is argued to reduce the cost to consumers of re-dispatches compared with our present-day decentralised approach. In short, it means that generators would no longer be eligible for compensation if network constraints are such that they can’t be dispatched.
The cost of re-dispatches of generation due to network constraints has been rising over recent years and was around £1.1 billion in 2021-22. It might be noted, though, that the total constraint cost includes both compensation for curtailment of plant in exporting areas of the network and the cost of production – at present using unabated burning of natural gas – in importing areas. One presumes that the latter cost would also be incurred in a centralised, LMP dispatch if the generation background and network were as they are today.
Differences in energy prices location by location are argued by proponents of LMP to incentivise investment in new generation, demand and dedicated providers of ‘flexibility’ in locations that are useful to the system. However, particularly for generation, it is hard to see this creating positive differences over the critical investment period of the coming decade. This is because generation investment depends on many factors: yes, a credible business case with relatively high confidence in a revenue stream over years or decades, but also planning permission – heavily dependent not only on local environmental impact but also politics – resource availability, supply chains, and a healthy project pipeline. Each of these factors needs to be in place to deliver a project, something that takes many years.
There is a reason why the key network bottlenecks today are where they are: the best areas for development of wind generation lie behind them. A move to LMP would likely weaken one of these key investment factors – a credible business case – for much of the current generation pipeline. Wind capacity in Scotland is the most obvious example.
It is argued LMP could instead support an improvement in the business case for investment in alternative locations such as wind capacity in the south of England. However, if development of projects in these other locations is not already well advanced, they are unlikely to be ready for the 2035 target of a decarbonised electricity system. Moreover, as can be seen from the Electricity System Operator’s ‘holistic network design’, many sites in the south also depend on major network investment. In other words, if the network is not significantly developed then, before long, these, too, will become caught behind network bottlenecks. This points towards a second reason why network constraints today are as they are: delays in delivery of enhanced transmission network capacity. In this context, the example of one of the places most often cited as providing evidence of the success of LMP – Texas – is relevant. There, nodal pricing was introduced only in concert with centrally planned strategic development of the transmission network that was seen as essential to allowing wind to connect in the best locations.
In timescales to 2035, LMP in Britain risks penalising rather than incentivising investment in renewable assets while the sector catches up on transmission investment. The potential for LMP to positively influence siting decisions in the longer term to 2050 is more plausible, but still questioned by some.
There is near universal acceptance of the importance of ‘flexibility’ in continued decarbonisation, by which it is normally meant that we need flexible demand and storage to help to manage the variability of demand and power from renewables and balance the system. There are also outline descriptions of a mechanism by which LMP could provide improved revenue streams for flexibility which could, it is argued, support their efficient location and operation. However, we found very little analysis of how much of an impact these proposed mechanisms would have on investment decisions. Whilst we accept that the theoretical model proposed could have some impact, we are yet to be convinced that it would be the game-changer that it is claimed to be.
California provides an example of an LMP system that has seen significant growth in ‘grid scale’ battery storage in recent years – around 3.5 GW in the two years since June 2020. However, it is not clear that this has been driven by LMP-derived energy prices or, alternatively, whether it is simply a counterpart to the growth in solar and something that would have happened with other market frameworks in place. In fact, modelling by the CAISO Market Monitor showed that under 2021 market conditions, only around 11% of battery revenue on average came from energy trading whilst 89% came from the provision of ancillary services (such as frequency regulation). So, whilst batteries are likely to carry out arbitrage once built, it appears, at least in the California case up to now, that it is non-energy revenue streams that are driving the investment.
Similarly in GB, the absence of LMP is not holding back investment in flexible assets like battery storage where a strong business case based on provision of various ancillary services exists already. Much has been made over recent years of the relative lack of participation of demand in ancillary service markets and of deliberate time-shifting of demand. Is that because of a lack of wholesale market signals or encouragement by the system operator? Or is it primarily due to a lack of half-hourly metering, innovation by Suppliers or insufficient electricity demand that is truly flexible such as electric vehicle charging?
We should expect the volume of flexible assets to increase even without wholesale market reform. This may lead to ancillary service markets becoming saturated. The locational value of flexibility would then become a more important investment driver for new assets, recognising that, for example, the times of surpluses and deficits of power relative to demand will be different for places dominated by wind compared with those dominated by solar. However, it would also be important to consider the extent to which limited transmission capacity can underpin flexibility business models. For example, periods of constraints out of Scotland are likely to last significantly longer than the discharge time of batteries, limiting their ability to cost-effectively reduce curtailment.
There is agreement on the centrality of flexibility to decarbonising power. The current lack of understanding of the likely shape of the business case for different types of flexibility in future is, however, a major evidence gap. It is one that must be filled in order to make good market reform decisions.
In making judgments about market reform we should not forget our overarching societal objectives and, in particular, the question of how serious we are about decarbonising electricity by the mid-2030s. There is a surprising amount of woolliness around the UK Government’s commitment to “fully decarbonise electricity by 2035, subject to security of supply”. The second clause in that statement appears to be interpreted in different ways across the industry. Our understanding is that it would be acceptable for unabated gas capacity to remain as a strategic reserve for emergency use, or even for a certain amount of capacity to generate for a handful of hours per year. But, in terms of energy supplied, the contribution from unabated fossil fuels has to be negligible.
If we are truly serious about fully decarbonising electricity by 2035 then the investment challenge across multiple asset classes dwarfs everything else. This includes investment in (a) renewable generation capacity; (b) schedulable low carbon generation, potentially including nuclear, gas and biomass with CCS, and hydrogen; (c) flexibility in the form of both short and long duration storage and flexible demand; and (d) transmission network capacity.
If the pathway to reform were simple, if an ‘off the shelf’ LMP framework could be introduced, if there were no major evidence gaps, and if investors were confident, we could get on with introducing LMP and have confidence in meeting our 2035 target. Unfortunately, none of those things are true.
The discussion above is very generation and flexibility-centric and might be argued to place the consumer insufficiently strongly at the centre of the debate.
It is essential that energy users get good value from their electricity system: a reliable supply, at low cost. With the massive reduction in the cost of production of electricity from renewables in recent years – it is lower than the levelised cost of energy from new gas plant, even before the price rises of last year or two – it is reasonable to expect that the cost of electricity to consumers in future will be lower than it is today, provided we get the market design right and keep ‘system costs’ to a minimum.
It is important not to just wave away the very big estimates of consumer savings, in the range of tens of billions of pounds over the decade or so after implementation, quoted in some recent assessments of the potential benefits of introducing LMP in Britain. However, as well as the nervousness we have around some of the assumptions used (such as the cost of capital, discussed above), we believe that the biggest potential gains for consumers in the next decade will come from increased use of renewables, to both reduce energy bills and better protect us from geopolitical uncertainty. We worry that this might be put at risk if market reform leads to an investment hiatus.
What emerges from our research is that, like the implications for flexibility, the way in which an LMP market in Britain would impact consumers is, as yet, poorly understood. In fact, the key options are yet to be properly articulated. In most US LMP markets, for example, the demand side is only exposed to locational price difference at the zonal rather than nodal level. However, some systems, such as New Zealand, do apply nodal prices to the demand side.
Then there is the role of Suppliers and whether they could be charged for demand on a nodal basis but regulated in the degree to which time and locational variation can be passed through to end consumers. The prospect of highly variable energy prices for consumers based on a postcode lottery has the potential to be a political hot potato. Attempts to avoid that by shielding certain consumers (e.g. households or businesses below a certain scale) from the locational price differences of an LMP market would also draw ire from those that miss out on lower prices – why can’t I use cheap, local wind energy to charge my car, power my heat pump or drive my manufacturing process? – and reduce the potential benefits that LMP could deliver.
The main conclusion that we reach from our research is that LMP is not, on its own, ‘the right answer’ for GB and a largely decarbonised power system by 2035. However, it might be part of the right answer, perhaps over a longer timeframe. The need for significant investment – and the potentially very large impact of rises in cost of capital on total cost – means that a package of complementary market reforms (which might or might not include LMP) would need to be designed very carefully, based on good evidence.
The very real concerns that we raise over the introduction of LMPs aren’t meant to imply that alternative routes would be easy. It’s been beyond the scope of our recent study to assess a variety of means of addressing potential misalignments between current GB market arrangements and what’s necessary for the future zero (or nearly zero) carbon system, such as whether decentralised scheduling of resources will make best use of storage and flexible demand, or how to get sufficient capacity of long-term storage located in the right places and ensure that there’s enough energy in it. But, certainly, that work needs to be done and conclusions reached.
If we are serious about electricity decarbonisation by the middle of the 2030s then our focus needs to be unerringly on delivery. We don’t have many investment cycles left before 2035 for big infrastructure: the large offshore wind farms, the big schedulable low carbon power stations and the new transmission network links needed. Our view is that we have to take a pragmatic approach to all elements of electricity system reform (including, but not limited to, electricity market reform) if we are going to be successful in delivering our objectives. We need sufficiently good – not necessarily theoretically perfect – ways of trading energy. We also need sufficient network investment, ways of meeting demand during periods of low wind output – the so-called ‘Dunkelflaute’ – and complete confidence that we know how to operate a power system that will be very different from today, dominated for many hours of the year by renewables and technologies based on power electronics, things that, in combination, we currently don’t know enough about.
About the authors:
Simon Gill runs the independent consultancy The Energy Landscape and is an associate of Regen.
Keith Bell holds the ScottishPower Chair in Future Power Systems at the University of Strathclyde and leads the UKERC research theme on Energy Infrastructure Transition.
Callum MacIver is a Research Fellow in the Department of Electronic and Electrical Engineering at the University of Strathclyde and a contributor to the UKERC research theme on Energy Infrastructure Transition.
The authors have worked together on a review of evidence and perspectives on what Locational Marginal Pricing would mean for the British electricity sector. Two reports have been published – a main document and a record of findings from stakeholder engagement.
This project has been supported by the University of Strathclyde’s Scottish Low Carbon Power and Energy Partnership, funded by SSE and ScottishPower. However, Gill, MacIver and Bell have full editorial independence and take full responsibility for any errors and omissions.