Billing me softly

Less than a month after Britain’s new Prime Minister took office, and despite a day one decision to cut VAT from electricity bills, yet another escalation in the war in Iran means that energy bills have not left the political agenda. As of early August, wholesale gas prices have surged back to around the highest levels since fighting began.

But the past four years have exposed a gap in the state’s toolkit: beyond the benefit system, the Government has no way of targeting support for households struggling with energy bills. When prices spiked in 2022, ministers fell back on universal support, at a cost of £44 billion. The new Government should not risk repeating previous mistakes: it should build a targeted scheme now, ready to apply a discount to gas and electricity unit prices if needed in early 2027. It does not make sense to rely on further universal support, where the richer half of the population get more than half the benefit.

But targeting is harder than it sounds. Passporting support to means-tested benefit recipients, as the Warm Home Discount does, reaches only around a quarter of British households, and would miss most below-average-income households. Nor is hardship confined to the very poorest: in 2022-23, more than 15 per cent of households in the second income quintile could not afford to keep their home warm.

To strike a balance between effective targeting and deliverability, we propose extending support to households passing an income test based on the highest income of the resident individuals. A threshold of £24,000 per year would reach around 40 per cent of households, and cover 75 per cent of those in the poorest four income deciles. A £2 billion scheme could provide an average of £175 per eligible household, or be tiered – paying around £220 to those with incomes below £18,000, and £85 above – to favour the poorest. Automatic enrolment via HMRC data should be the destination; for this winter, self-declaration verified by HMRC after the fact is imperfect but deliverable.

The new Prime Minister should be prepared to act on energy bills

Our new Prime Minister has a full in-tray, but the cost of living is certain to be near the top. When campaigning to be the Labour leader, he said that “Britain is paying too much for the basics” and has already announced a 6-month zero-rating of VAT on electricity bills to give households some “breathing space”.

But now conflict in the Middle East is re-escalating, and an apparent breakdown of the fragile peace deal alongside an increase in rhetoric have pushed wholesale gas prices back to around 150 pence per therm. This is close to levels seen shortly after the first attacks, which sent shockwaves through global commodity markets, and double the prices seen a year ago. The Ofgem price cap has already increased by 13 per cent from Q2 to Q3 this year, and although forecasts for the next (Q4) Ofgem price cap show expectations of just a small uptick, the risk of a large jump in prices in Q1 2027 is very real.

The VAT cut, coming into effect from October 1, and worth around £45 per year to the ‘typical’ household, will take some of the sting away from these jumps. Shifting taxes away from electricity is a welcome step towards rebalancing fuel prices to aid the net zero transition, but does little for households worrying about gas heating costs as colder nights draw in – the dominant heating source in Britain, with 84 per cent of homes on the gas grid.

As such, new (and returning) Ministers should look to complete the job that previous governments left undone: setting up a system that can be easily turned on to shield vulnerable households from high energy bills, rather than having to resort to providing costly universal support, as happened in 2022, or risk months of hardship. This should be done ahead of the Q1 2027 price cap, meaning there is pressing work to do now on scheme design, data linking and administrative infrastructure. So, what should that contingency plan look like? [1]

Emergency energy support should be targeted and temporary

We have said in the past that a scheme should be targeted and temporary, and the former Chancellor said similar.

By “temporary”, we mean that any single round of support should be time-limited. The natural trigger for activation is the energy price cap itself. Its quarterly reset provides a window for quick intervention when prices are high (i.e. a high price cap) or when bills are high (i.e. during the winter), and we discuss later how we think this could be operationalised. Crucially, though, “temporary” should not be seen as meaning “one off”. On the contrary, the past few years have shown us that the UK government needs to develop the administrative infrastructure necessary to deliver support and then keep that system ready and waiting for future shocks, rather than be bounced into announcing poorly designed schemes each time.

By “targeted”, we mean that support should be directed to households that need it. The alternative is either no help, or falling back on universal approaches that cannot distinguish struggling households from those for whom higher energy prices are inconvenient but far from ruinous. Untargeted schemes are expensive, far from ideal for a new Government facing the same (or even greater) fiscal pressures than its predecessor.

At present, three approaches are prominently discussed.

One is to remove the remaining ‘levies’ from electricity bills, transferring around £2.6 billion of costs from billpayers to taxpayers and thereby finishing the job that Rachel Reeves started in the 2025 Budget. In the long run, such a switch would be very welcome: the continued use of energy bills as a secret tax and benefit system leads to unwelcome distributional consequences, and the loading of costs onto electricity but not onto gas actively works against the net zero agenda. So this is the right thing to do, but, just like the VAT cut, is not a substitute for emergency support designed to prevent hardship. Levy-free electricity would be less beneficial to households shivering through the winter months than action on both gas and electricity (especially as three quarters of the typical family’s gas use takes place between October and March).

The second is a form of targeting – providing an ‘’essential block” of cheap or free energy for all households. But were the same block offered to all households, it would function as a flat cash transfer – effectively a Warm Homes Discount for everyone – and one that risks being unfair to those whose essential use is underestimated by the scheme’s parameters. Proposals to circumvent this by offering larger allowances to families with children or a disabled member would require gathering details from all 28 million households and adjusting metered rates accordingly, a task sufficiently complicated that it might as well be used to means-test them instead. Block schemes also do not help the net zero transition: unlike a unit-rate discount, they do nothing to make the additional electricity needed to run a heat pump or electric vehicle cheaper.

The third is one broadly in line with that employed in 2022: guaranteeing that prices will not breach a certain level, with protection applying either to the total bill or to a part of it. A new Government grappling with the same fiscal pressures as the last cannot afford to distinguish so poorly between households that are struggling and those who are not, and should certainly not open itself to the huge risk of running a short position against perennially volatile global energy markets.

A better approach would be to reduce the unit prices of gas and electricity for certain households. Making each marginal unit cheaper would, by its very nature, see those with higher levels of consumption receive more protection. Metered prices are relatively straightforward to change, and indeed are altered each time a customer changes tariff or the price cap renews. The main task, then, is designing how support should be targeted, and how this should be assessed. We discuss this below.

Relying on passporting support via the benefit system is not enough

One way to target support would be to limit it to those receiving means-tested benefits, building on the infrastructure used to deliver the Warm Home Discount. For that scheme, the Department for Work and Pensions (DWP) shares information on benefit receipt with energy companies, who then administer the discount as a deduction on electricity bills between October and March, or as credit for those on prepayment meters.[2]

An advantage of passporting is that households do not need to apply, significantly reducing their administrative burden and ensuring no one who is entitled misses out. But the major disadvantage is that passporting receipt of means-tested benefits reaches only around a quarter of households. Indeed, the majority of households in the bottom half of the income distribution would not receive help, owing either to incomplete take-up or a lack of eligibility, as Figure 1 shows.

Figure 1: Passporting support via receipt of means-tested benefits misses the majority of poor households

An option to reach more below-average-income households would be to also include those in receipt of non-means-tested disability benefits. This would extend help to around 9 per cent more households in the bottom half of the income distribution, but also to around 10 per cent more households in the top half. It is therefore not the most cost-effective way of granting eligibility to more poor households.

But evidence suggests it is not only the very poorest who have struggled to pay their bills in recent shocks: Figure 2 shows how the proportion of households saying they cannot afford to keep their home warm enough rose most sharply for the two poorest quintiles between 2020-21 and 2022-23. For the bottom quintile this is perhaps to be expected, as they have consistently reported the highest rates of being unable to afford adequate warmth, and their very low incomes leave them much less of a cushion to absorb rising prices.

More striking is the change for the second quintile. In 2020-21, the proportion of these households unable to afford adequate warmth sat just 1 percentage point above the third and 3 percentage points above the fourth and fifth quintiles. By the 2022-23 energy shock, those gaps had widened to 4 and 9 percentage points respectively, with more than 15 per cent of second quintile households reporting that they are unable to afford adequate heating.

Figure 2: Households in the bottom two income quintiles are most likely to be unable to afford high energy bills

This suggests that the Government should aim for a method that targets a greater proportion of households than can be achieved from a simple passporting of means-tested benefits. There is, of course, no exact science in determining how high up the income distribution energy support should go (and the answer likely depends on the size of the energy shock), so the Government should develop a mechanism whose reach can be scaled up or down accordingly.[3]

Targeting based on the highest-income individual in the household strikes a balance between fairness and deliverability

To efficiently and effectively target more households than those receiving means-tested benefits, we propose that support be paid automatically to families where someone receives a means-tested benefit but also to households meeting an additional income test.

But there are many ways to measure income, so which is the best to use?

The fairest way of determining need for most households would be to use a threshold that depends on a household’s total income. But it is worth being clear why the household is the relevant unit at all. Energy is billed at the meter level, and each property typically has one meter (per fuel): a unit-rate discount therefore applies to a whole property’s consumption.[4]

In most cases a property will contain a single household, but households often contain more than one family (i.e. multiple benefit units) – adult children living with parents, say, or friends sharing a house. So, while the benefit system assesses income at the benefit unit level, a discount delivered through energy bills should be assessed at the household level. This is where a total household income test runs into trouble. It would be difficult to administer where multiple benefit units share a household, and complicated to understand: even in the social security system, benefit units living together do not normally have to share financial information with each other. We think this principle should extend to energy support, for reasons of practicality as much as privacy.

An alternative to using total household income would be to target support based on the highest individual income in the household. This is similar to the way in which the Tax-Free Childcare (TFC) system and the High Income Child Benefit Charge (HICBC) operate, although both of those look at the incomes of the adults in a one- or two-adult family (i.e. benefit unit), and not the whole household. This would, of course, still require a degree of information sharing between household members, but it would be simpler and less susceptible to error than asking individuals to calculate their total household income.

Fortunately, the distributional difference between targeting support based on total household income and the highest income in the household is small. Figure 3 shows illustrative options for targeting support to households if their total (equivalised) annual household income is below £30,000, or if the annual income of the highest income individual is below £24,000; both targeting around 40 per cent of households. 4F[5] Under the total income option, 78 per cent of households in the poorest four income deciles would be eligible for support, compared with 75 per cent under the individual income option. [6]

Figure 3: Targeting based on the highest income in the household has a similar distributional impact to targeting total household income

However, just like the TFC and HICBC, such a system would lead to some rough justice: households with a single high earner above the threshold would not be eligible, while those with multiple earners each just below the threshold would be. This single-earner penalty has made the HICBC a frequent source of complaint, so ministers should adopt the approach with eyes open to its drawbacks (although this anomaly is arguably easier to justify for time-limited support than for a standing tax charge).

Figure 4 shows that the main group who lose out (proportionally) from assessing based on a highest individual income, rather than total household income, are working-age couples with children, who are more likely to have a single higher earner than other family types. Of 1.8 million such families eligible under our hypothetical total income option, 490,000 would not be eligible under an individual income option, and 71 per cent of these are in the poorest fifth of households. Conversely, the main beneficiaries of assessing based on the highest individual income rather than total household income would be pensioner households: 780,000 would become eligible, of which 81 per cent (630,000) are in income decile five or above.

Figure 4: Targeting based on the highest income in the household rather than total household income disfavours working-age couples with children, and favours pensioner families

This is an uncomfortable trade-off, and one worth stating plainly: it withdraws support from a group of low-income families with children to gain administrative simplicity, while extending eligibility to pensioner households that are, on average, less likely to sit in the poorest deciles. On balance, though, we still favour targeting support based on the highest income in the household, as the alternative of requiring households to calculate their total income risks the kind of errors and delays that could leave many more without timely support. But the cost to poorer families with children is real, and the Government could consider mitigations – such as a more generous individual income threshold for households with children – to soften it.

Assessing income over the three months before the start of a scheme strikes the best balance between deliverability and determining current need

The next question would be over what period income should be assessed. For a scheme that is implemented for a quarter of a year at a time (i.e. coinciding with the setting of the energy price cap), we propose assessing income over the 3 months immediately before the point at which households would need to apply. This would be a reasonable proxy of current need, and is a long enough period to mitigate the issues with misaligned assessment periods and earnings cycles that arise from Universal Credit’s monthly assessment. It also avoids some disadvantages of other options. For example:

• Assessing income on an ongoing basis throughout the scheme, while probably the best indicator of current need, would be administratively difficult and costly.
• Assessing expected income over the length of the scheme (like Tax-Free Childcare) would lead to errors that would need to be corrected after the fact.
• Assessing based on the most recent financial year for which HMRC has complete self-assessment records would enable accurate assessment of income from sources like self-employment, property and investment, but would not be a good indicator of current need (for example, a scheme implemented from Q1 2027 would have to use income from the 2024-25 financial year).

A tiered discount could help target more support to poorer households

Once it has determined the concept of income used to determine who receives support, the Government must then decide when support should kick in and the level of discount that should be provided.

One key choice is whether discount levels are flat across eligible households or tiered to give poorer households higher discounts. For example, for a scheme costing £2 billion (a cost that should be covered from public funds rather than recouped via yet another surcharge on energy bills), a flat discount across all households with highest individual income below £24,000 would be worth, on average, £175 per household. But a two-tiered system costing the same could grant support of around £220 for the average household with highest individual income below £18,000 and around £85 for households with highest individual income between £18,000 and £24,000. However, introducing a second income threshold would come with increased administrative complexity, which the Government must weigh against its advantages for targeting poor households.

At least for this winter, non-benefit recipients will need to declare their income

The final question facing the Government when designing a Targeted Energy Discount scheme is how eligible households will be identified.

The longer-term goal should be automatic enrolment. This would avoid issues of incomplete take-up that are common among social tariffs for water and telephony, for example, and reduce the administrative burden for households (and hopefully lessen it for the Government too). Legal gateways exist for the potential use of HMRC tax data to target energy support without primary legislation,[7] but implementing an automatic scheme would require significant work on data-linking infrastructure between HMRC and energy companies.

Additionally, it is not clear how non-wage income could be assessed in real time for an auto-enrolled scheme: HMRC Real-time Information (RTI) data does not currently contain information on income from self-employment, rental properties, investment and savings, and dividends. Some of this income is not known until individuals finalise their self-assessment forms, so, as we said earlier, for a hypothetical scheme operating in early 2027, data on all sources of taxable income would be known only up to 2024-25.[8]

For this winter at least, then, it seems very likely that non-benefit recipients would need to declare their income. This could be done fairly simply by mirroring the way that eligibility for Tax-Free Childcare is currently assessed: our proposal is that eligible households tick a box in an online account to declare that the income of the highest-income individual is below the relevant threshold, and HMRC then verifies this after the fact when it has data on all forms of income. Such an approach still has drawbacks: it would require a significant amount of work to set up in time as the application and assessment infrastructure would need to be built and the exact parameters of the scheme would need to be decided on. Inaccuracies in HMRC’s address records would limit how accurately income can be verified using RTI data, and this is unlikely to be fixable in a matter of months. Self-assessment information cannot be assessed until at least a year after a scheme has ended. And there would be higher levels of fraud and error and lower take up than with auto-enrolment, as households could misrepresent or miscalculate their income, and receipt would depend on knowing about the scheme. Despite these, it is still likely the most deliverable option in the short term that extends support beyond those receiving means-tested benefits and without the waste associated with a universal system.

Conclusion: the Government can’t afford to risk not having an effective scheme ready to go at the start of next year

Energy prices have spiked twice in the past four years, and there is a real risk of a further rise in January 2027 should the situation in Iran deteriorate. The need to have a better system for intervening in household energy costs is therefore unlikely to go away.

A key lesson from both the Covid-19 pandemic and the war in Ukraine is that the UK government lacks any effective means of targeting support at households beyond the benefit system. It should correct this now and build a targeted scheme that can be switched on when needed, reducing unit rates to help those who struggle with high energy costs but fall outside the social security system.

Such a scheme should be ready to protect vulnerable households this winter, so as to reduce the need to fall back on expensive universal support. We propose introducing an eligibility threshold for the highest individual income in a household alongside passporting off means-tested benefit receipt, assessing income over the three months before the start of the scheme and, for this winter at least, requiring non-benefit-receiving households to declare their eligibility. This scheme would not be perfect: the inability to auto-enrol those who do not receive benefits creates and administrative burden for households and potential take-up issues, and there would be some rough justice in using highest individual rather than household income. But the experience of the Covid pandemic shows us that governments can move quickly to devise new programmes when they need to, and the scheme we have outlined should be deliverable this winter, where it would certainly be superior to poorly targeted and expensive universal options.

[1] This note refers to household energy bills in Great Britain; bills in Northern Ireland are regulated separately and are not currently subject to Ofgem’s price cap.

[2] Warm Home Discount is an annual £150 discount on electricity bills, made in the winter months. Families get the discount if they receive Pension Credit, Housing Benefit, income-related Employment and Support Allowance, income-based Jobseekers Allowance, Income Support or Universal Credit and their energy supplier is part of the scheme.

[3] One important caveat is that discounted energy bills can only reach those with a domestic supply contract and with a meter capable of passing on a rate change. This means that households with traditional (non-smart) prepayment meters would need supplier or customer action to receive the discount, while those on heat networks or tenants paying for energy through their rent could not be reached at all.

[4] Strictly, the scheme would apply to meter points rather than households: these usually align one-to-one, but not always – houses in multiple occupation may share a single meter across several households (with residents often paying for energy through rent), while a small number of properties have more than one meter. Where multiple households sit behind one meter, all would benefit from a discount triggered by any one of them, as is already the case for the Warm Home Discount.

[5] Taxable income is used. This is equivalised using the OECD before-housing-costs equivalisation scale for the household income option as a way of implementing different thresholds for different household sizes.

[6] Income quantiles are calculated based on net equivalised after-housing-costs household income, the standard measure for assessing the distribution of income (and that used in DWP’s Households Below Average Income), but is different from the taxable income measure we propose to use for targeting energy support, which is why coverage is not 100 per cent even in the poorest income deciles.

[7] The Digital Economy Act 2017 created a framework of public service delivery powers that allow specified public authorities to share data for defined purposes, one of which is reducing and preventing fuel poverty. Provision is made in the Act to allow government departments and energy companies to share relevant data with each other.

[8] A potential workaround could be to let households who are ineligible thanks to unearned income opt out of the scheme, and for those that do not to have the amount of the discount taxed back in the following financial year (as is now done with Winter Fuel Payments).

Data citations: Family Resources Survey: Department for Work and Pensions, NatCen Social Research. (2021). Family Resources Survey. [data series]. 4th Release. UK Data Service. SN: 200017, DOI: http://doi.org/10.5255/UKDA-Series-200017; Households Below Average Income: Department for Work and Pensions. (2021). Households Below Average Income. [data series]. 3rd Release. UK Data Service. SN: 2000022, DOI: http://doi.org/10.5255/UKDA-Series-2000022