Why this recommendation matters
Flexible resources need clear information on which to base the use of their flexibility. For a minority of large, sophisticated actors that information can come from direct market participation, where markets exist. For most resources, from factories to households to the devices within them, the information will most likely have to come from the tariffs they pay or the programmes they enrol in. These are the instruments through which millions of independent decisions can be steered. What they coordinate is the system’s energy balance itself: the volumes consumed, shifted and injected hour by hour. In liberalised systems, this is achieved through day-ahead and intraday trading and imbalance markets. Elsewhere, tariffs and programmes have to be designed to approximate the time- and location-varying costs of balancing supply and demand.
Today’s prices and tariffs often communicate inaccurate information about the state of the power system. Insufficiently dynamic energy tariffs fail to differentiate between midday solar surplus and evening scarcity, or they reflect past system conditions, e.g. when night-time tariffs are lower than daytime tariffs. Volumetric grid charges fail to communicate the fact that the grid has to be designed for peak use, resulting in significant fixed costs. Where they are combined with net metering or net billing arrangements, they can also shift grid costs from consumers with distributed resources, who are generally wealthier, onto other consumers without the financial resources or ability to install them.
Existing price signals often also stop at a threshold, so the devices best suited to automated response – EV chargers, heat pumps and batteries – sit behind meters that price them like the rest of the household. This signal problem is distinct from the participation problem: even when these resources are eligible to participate, they are not exposed to a clear and accurate signal to respond to.
What coordination requires is granular information through prices or price-like signals that show with growing temporal and locational resolution when and where electricity is scarce or abundant. Markets produce this information through day-ahead, intraday and balancing prices. Where no such market exists, the information has to be constructed: the shadow prices of a centralised dispatch offer an approximation, but one that typically captures the variable costs of generators while missing the price effects of storage, demand response and scarcity, so usable reference prices have to be built deliberately. This information then needs to reach users. This can be done via at least three routes: via direct exposure to market prices, often suitable for large and sophisticated actors; via dynamic tariffs set by retailers, or by the utility where retail is not subject to competition, for smaller users; and (where used) via activation signals in direct-control programmes and virtual power plants, tied as closely as practicable to the real-time supply–demand balance.
How to act on this recommendation
Stage 1
Introduce the signals that work with today’s metering and institutions, and reach the largest resources first.
Time-variable energy tariffs should be introduced to give consumers a reason to move consumption towards cheap, clean hours. Asset-specific tariffs should extend stronger signals to high-flexibility devices, pricing the device rather than the household wherever separate or embedded metering allows. Grid tariffs should become time- and capacity-differentiated, beginning with capacity-based tariffs that need no smart metering. Large consumers above defined thresholds should face dynamic price signals through market participation, dynamic contracts or pass-through of utility marginal costs where no market exists. Where no wholesale market operates, the utility or system operator should construct and publish the time-varying reference prices these signals draw on. Direct-control and interruptibility programmes can give the operator a firm contracted call on large flexible loads. These measures connect to the metering in recommendation 5.2 only where their design requires it, which is what makes this stage broadly available.
Stage 2
Extend dynamic signals to willing retail customers and extend control of distributed assets.
Offer dynamic pricing to retail customers on an opt-in basis, retaining protective defaults and realising the value through automation. Direct coordination of distributed assets, through virtual power plants and comparable arrangements, delivers firm response from small resources where prices alone would not, and depends on the observability and device-embedded metering in recommendation 5.2 and the data access rules in recommendation 7.3. Reform prosumer compensation towards remuneration at system value. And introduce protections to ensure that customers who cannot participate are not left carrying a disproportionate share of system costs.
Stage 3
Add locational dimensions to the coordination signals.
Locational signals, through nodal or zonal pricing and locational grid charges, help steer where assets choose to locate and how they operate. Generation typically concentrates in resource-rich areas and demand in electrification clusters. This action is deliberately last: the simpler locational instruments arrive earlier through the connection charges and siting rewards in objective 4, the hosting-capacity maps in recommendation 4.2 and the local service procurement in recommendation 6.3, while full locational pricing depends on granular metering and settlement in recommendation 5.2, governed data access in recommendation 7.3, and substantial institutional capacity.
Core measures
Implementation considerations
Stakeholder interests can diverge on how the value generated by more granular coordination signals is distributed. Granular coordination signals are efficient for the system as a whole, but their introduction often creates winners and losers relative to the status quo. The winners are those who can be flexible, and those in parts of the system where activated flexibility lowers prices and tariffs. The losers are those whose bills rise against the status quo, and those who cannot or do not wish to adjust. That group includes vulnerable consumers, and also businesses who are reluctant to spend management attention on becoming flexible and are wary of the effect on their competitiveness. A second divide concerns who keeps the value. Intermediaries, including utilities and the operators of virtual power plants, do not automatically pass savings on to the consumers whose flexibility produces them. Retailers have stronger incentives to pass value through, but retail is not open to competition everywhere. Regulators should watch the points in the chain where value can be retained. Building an ecosystem in which resources respond smoothly to prices and tariffs is also itself a complex undertaking, although it can be an advantage that this complexity falls on retailers, aggregators and automation providers rather than on the system operator.
Equity considerations centre on the same distributional effects, and most of all on protecting vulnerable consumers. The ability to respond to signals is unevenly distributed, correlating with home ownership, capital, automation and digital engagement. Locational signals add a second axis, since they can disadvantage households and businesses simply for their location, such as in rural or grid-constrained regions. Safeguards should keep locational charges from entrenching regional disadvantage. Design protections appropriately from the start: opt-in rather than forced exposure, allowing defaults under fair conditions, access to enabling technology and automation, and testing reforms for regressive effects before they are introduced. Prosumer compensation reform can be highly contested since it can change the economics of investments consumers have already made. Fair transition arrangements for early adopters can help make the reform politically sustainable. Automated response raises further consumer protection questions around data and privacy, addressed in recommendations 7.3 and 9.2.