Heat Network Accounts: Why the Deficit Matters More Than the Ofgem Deadline

Everyone in the sector is talking about registration. Almost nobody is talking about the fact that every month a heat network deficit sits unquantified, some of it becomes money you can never get back.

I have spent this year inside heat network accounts at a scale I did not expect. Thousands of connected units. Schemes ranging from a couple of dozen flats to one of more than a thousand apartments. Backlogs running four and five years deep. Data arriving from more than a dozen different metering and billing agents, no two of them formatted the same way.

What I did not anticipate is how consistently the same conversation happens. Someone shows me a set of draft accounts with a deficit on the balance sheet, described as recoverable from leaseholders. The figure is large, often larger than a full year of income. And when I ask when those costs were first demanded, or whether a holding notice was ever served, there is a pause.

That pause is the most expensive thing in the sector right now, and it has nothing to do with Ofgem registration.

What are Heat Network Accounts & How Do They Differ from Service Charge Accounts?

If you have never had to prepare one, the shape is unfamiliar even if you produce service charge accounts routinely.

A heat network is a small private utility inside a building. Rather than each flat having its own boiler and its own supplier contract, the building buys gas and electricity in bulk at commercial meters, generates heat and hot water in a central energy centre, and pipes it to each dwelling, where a sub-meter records what arrives. The resident buys heat from the landlord at a tariff, not gas from a supplier of their choosing. They cannot switch. That captive position is precisely why Parliament handed the sector to Ofgem.

The accounting consequence is specific. In a service charge account, income is the demands you raised, a clean figure from your own ledger. In a heat network account, income is consumption-based, calculated by a third-party billing agent from sub-meter reads, across separate charge types, adjusted for credits and re-bills through the year. You cannot lift it from anywhere. You have to rebuild it.

That rebuild is the engagement. On a 26-unit block over four years, it meant working through 36 monthly ledger tabs and just under 5,000 individual bill lines. On a thousand-unit scheme, scale accordingly.

Why Do Heat Networks Fall Into Deficit?

This is where people reach for the energy crisis as an explanation, and it is only half right. Wholesale prices were the trigger. The structure of heat network accounting is what turned a price shock into a permanent hole.

Four things compound, and they are worth separating because they have different fixes.

The tariff is set before the year begins. It is an estimate. When wholesale gas doubles in March, the network carries on collecting on last year's assumptions while paying this year's prices, and the correction only arrives at the next annual review, by which point twelve months of under-recovery is already banked. Residents behind a commercial bulk meter got none of the domestic price cap protection — they were covered instead by the non-domestic relief schemes, at a materially different level of support and with pass-through obligations of their own. That is why 2022 and 2023 were so brutal on communal heating specifically.

Energy bought never equals energy billed. The bulk meter measures what enters the plant. The sub-meters measure what reaches each flat. Between them sits generation inefficiency and standing losses from pipework that stays hot 24 hours a day, whether anyone is home or not. On many networks, consistent with DESNZ heat network efficiency benchmarking, a third or more of purchased energy never lands on a billable sub-meter, and on poorer-performing schemes the figure approaches half. A tariff that is perfectly correct per kilowatt hour still under-recovers in total unless those communal losses are funded somewhere.

The standing charge is usually too low. It has to carry maintenance, the billing agent's fee, metering, the energy centre's own power, and heat bleeding into empty units. Setting it properly is politically uncomfortable, so it tends not to happen.

Income demanded is not income received. Prepayment units pay up front, credit units are billed in arrears, and some do not pay at all. Voids still draw standing losses.

Set those four alongside each other and the deficit stops looking like an accident. It looks like an outcome.

What Does a Heat Network Deficit Look Like Over Four Years?

Take an illustrative scheme. The figures below are synthetic, built to demonstrate the mechanics, and deliberately not modelled on any account I have worked on150 units, year ending 31 March, four years outstanding, being prepared now.

The last largely pre-crisis year runs a modest surplus. Then the winter of 2022–23 arrives, the tariff is still set on 2021 assumptions, and the heat network takes a £167,000 hit in a single year. Tariffs catch up after that and the annual position improves steadily, but the ground is never recovered. Cumulative deficit £316,000, roughly 1.4 times a full year's income, on a mid-sized block.

Now the question that matters. How much of that £316,000 can actually be recovered from leaseholders?

Why Does the Tariff-Setting Model Widen the Gap in Heat Network Accounts?

There is a mechanical reason the tariff keeps falling behind, and it sits upstream of any single bad year. Heat network financial models, including the government's own template financial model used to appraise these schemes at the outset, typically escalate projected income and projected operating costs using separate indexation assumptions, often different indices moving at genuinely different rates, because that is the only honest way to forecast two things that behave differently over time. Nobody sets out to build in a gap. But once the two assumptions are set at financial close, they are rarely revisited, and if they turn out to diverge even modestly in practice, the effect compounds every single year rather than correcting itself.

Take the illustrative scheme again, picking up from the £232,000 income and £288,000 costs already shown for 2025. Project both forward four more years under two versions of the same assumption. In the first, income and costs are both escalated at 3% a year, roughly a fixed-index world where the original appraisal assumption turned out to hold. In the second, income still escalates at 3%, but real costs, energy, maintenance, the billing agent's own fee, run at 6%, a gap no wider than what many schemes have actually experienced since 2022.

A three-percentage-point gap between the two assumptions, sustained for just four years, adds roughly £94,000 to the deficit on this illustrative scheme, on top of everything already sitting in the table above. Nobody has to make a single billing error for that to happen. The tariff can be applied perfectly, every quarter, and the gap still opens, because the assumption feeding it was set once, years ago, and never checked against what actually happened in the heat network financial accounts. That is the specific thing worth asking of any scheme with a multi-year backlog: has anyone gone back and tested the index the tariff was escalated on against the index the real cost base actually moved on? On most schemes I have looked at, nobody has.

Do Heat Networks Pay Business Rates on Their Energy Centres?

Yes, and it is a cost line that rarely gets the scrutiny the rest of the heat network accounts get, largely because it is set centrally rather than negotiated, and because the mechanism is unfamiliar even to accountants who are otherwise comfortable with property.

Energy centres are typically valued for business rates on a contractor's basis rather than by comparison with rental evidence, the same valuation method used for specialist infrastructure like schools and hospitals. The valuation works through estimated replacement cost, adjustments for age and obsolescence, and the value of the site, to arrive at an effective capital value. Government then sets a decapitalisation rate centrally, which converts that effective capital value into a rateable value. For district heating undertakings that rate has held at 4.4% since 2017, and it has been confirmed as remaining unchanged for the 2026–29 rating list. The rateable value produced is then multiplied by a business rates multiplier to arrive at the actual annual bill, and this is the part that has genuinely changed: the flat, single-figure multiplier many older scheme models were built around was replaced from April 2026 with a banded structure, varying by the property's rateable value rather than one number applying to everyone.

Put a number on it. Take an illustrative energy centre for our 150-unit scheme, with an effective capital value for rating purposes of £2,000,000.

Just over forty-two thousand pounds a year, before the 1p transitional supplement that applies in 2026/27 to fund the transitional relief scheme (removed where transitional relief or Supporting Small Business Relief applies), which adds roughly £880 to this example. On a fairly modest energy centre, and that is before anyone has asked whether the figure in a given set of accounts is even using the current banding. Any scheme still running a rates line based on a multiplier assumption from before April 2026 is carrying a stale figure into a live cost, and it is worth an hour of anyone's time to check which band a specific energy centre now actually falls into against the current VOA figures, rather than trust an inherited spreadsheet built on the old flat rate. This is not one of the four causes of deficit set out earlier. It is a separate, additional cost line, and one more reason the "clean per-kilowatt-hour tariff" in the earlier section is not the same thing as an adequate one.

Are Heat Network Charges Covered by Section 20B?

If the heat charge is a service charge within the meaning of section 18 of the Landlord and Tenant Act 1985, then section 20B applies. Where relevant costs were incurred more than 18 months before a demand for payment is served, the tenant is not liable for so much of the service charge as reflects those costs in the heat network accounts, unless, within 18 months of the costs being incurred, the tenant was notified in writing that the costs had been incurred and that a contribution would subsequently be required.

Where the heat is not a service charge, Ofgem's back-billing condition applies instead, and it is a materially shorter window. So the recovery question turns entirely on a single characterisation: is this a service charge, or a separate commercial supply? That single question drives more consequences than anything else in this work, and it is answered from the lease, not from the accounts.

What is genuinely interesting is that the new regime does not sweep this distinction away. It builds it in. Ofgem's back-billing condition explicitly excludes amounts recovered through a service charge from its own 12-month limit, and states that section 20B (and any other relevant legislation) "will take precedence" for those cases, "until further notice." Ofgem has also confirmed it is working with the housing ministry on the interaction between the two regimes, and that a strict 12-month limit applied to service charge accounts would leave landlords no time to reconcile costs, prepare year-end accounts and issue balancing charges. Even in the service charge case, though, Ofgem encourages authorised persons to keep within the 12-month period wherever they can.

Two bodies of law, written independently, deliberately kept apart at the point they overlap. In the meantime, the practical answer is: establish the characterisation per scheme, from the lease, and record the clause references directly against the heat network accounts. Everything downstream follows from it.

How Does the 18-month Section 20B Rule Affect Heat Network Cost Recovery?

Read the worked example above against the statute. Section 20B operates on a simple test: were relevant costs demanded from the leaseholder within 18 months of being incurred, or was written notice of those costs and a future contribution given within that window? If neither happened, the leaseholder is not liable for that portion of the charge, full stop.

Only the year to March 2025 falls anywhere near an 18-month window measured from August 2026, and even then only its closing weeks: costs incurred from around February 2025 onward. Everything earlier in that year, and the £260,000 that accumulated before it, was incurred well outside the window. If those costs were never demanded, and no holding notice was served at the time, the leaseholders are not liable for them. Not "difficult to collect." Not liable.

That is the recoverability clock. It has been running the whole time, and in most portfolios nobody is watching it.

How Does Ofgem's Back-Billing Rule Apply to Heat Network Charges?

Where the heat charge sits outside the service charge regime, Ofgem's own condition applies directly: back-billing should be a last resort, and where it is necessary, it should be limited to 12 months from the date of the bill in the heat network accounts, to ensure all consumers are treated fairly.

The condition allows for a small number of consumer-fault exceptions, where the supplier may recover beyond the 12-month limit. These include where the consumer has obstructed a meter reading, has a responsibility to provide a reading and has not done so, or has failed to report or fix a fault within their control that resulted in unpaid consumption. Outside those exceptions, the limit holds regardless of how good the supplier's own reasons for the delay might be.

Ofgem's guidance also states plainly that networks already signed up to the Heat Trust's Scheme Rules, which have applied a 12-month back-billing limit since before statutory regulation existed, should simply continue to meet that standard. For anyone already Heat Trust registered, this condition changes very little in practice. For anyone who is not, it is a new hard limit with real teeth behind it.

How Should Heat Network Billing Data be Reconciled?

I want to be concrete here, because "we reconcile the billing data" covers an enormous range of rigour in heat network account reconciliation. There are six links, and each answers a different question.

Meter to consumption. Were the reads used for billing actual reads, and are they complete for every unit and every period? Estimated reads are not the same quality of evidence as actual reads, and the difference should be visible in the file rather than smoothed over. Ofgem now expects metered bills to rest on an actual reading at least annually.

Consumption to income demanded. Was the right tariff applied to the right consumption for the right period, including every credit and re-bill? Credits move income between periods and can distort a single year badly.

Income demanded to the agent's reported total. This is the control I would protect above all others. Rebuild income independently from bill-line detail, then agree it to the billing agent's own reported figure. Two independent routes to the same number. When they agree, the income figure is evidenced. When they do not, you have found something real, and finding it is the whole point. Do not adjust your rebuild to match the agent. Investigate, document, report.

Income to cash. What was actually collected and remitted, against what was demanded.

Cash to closing arrears. Opening arrears plus income less cash equals closing arrears, or you have a problem.

Bulk energy to billed energy. The system efficiency reconciliation. This is a materially larger piece of work and needs bulk supplier invoices and bulk meter reads, so treat it as separate scope. It is also becoming unavoidable: the authorisation conditions require heat supply contracts to disclose key performance indicators of the network including network efficiency. Somebody has to calculate that number.

If a set of accounts cannot show you the third and the fifth of those, I would not rely on the income figure.

What Are the Most Common Mistakes in Heat Network Accounting?

Energy accruals on multi-year catch-ups. Bulk energy invoices are frequently estimated and trued up much later, sometimes years later. On a four-year reconstruction this is the single largest source of misstatement I see, and it is easy to miss because the arithmetic all ties. The same pressure point shows up in how these costs get modelled from the outset. HNDU's own template financial model treats operating cost inputs for generation, distribution and retail as summary-level entries indexed forward year on year, which is fine for appraisal purposes but leaves plenty of room for a real invoice, arriving late and estimated, to sit undetected against a modelled assumption.

Inter-fund borrowing, undisclosed. When a network runs out of cash it borrows from the service charge fund to pay the energy suppliers. That is common and often unavoidable. But service charge monies are trust monies under section 42 of the Landlord and Tenant Act 1987, and ICAEW TECH 03/11 treats using the funds of one property to meet the costs of another as a breach of trust. Borrowing between the heat account and the service charge account of the same property is a softer question, but it still depends on the lease permitting it and it should be disclosed plainly rather than netted into a balance sheet caption.

VAT that quietly changes. I have seen VAT charged on heat income up to a point and nil afterwards, with nothing on file explaining the switch. The likely story is a move from treating the heat as a separate metered supply to treating it as part of a mandatory residential service charge, but that is a hypothesis, not a conclusion, and the two are taxed very differently.

Which VAT Rate Applies to Heat Network Charges?

A supply of heat is a supply of goods for VAT purposes, standard-rated unless it qualifies for the reduced rate under Schedule 7A VATA 1994 for qualifying use. HMRC's guidance in VAT Notice 701/19 confirms that supplies of steam and heated water for qualifying use are reduced-rated with no de minimis, that air-conditioning and ventilation from central plant follow the same treatment for qualifying use, and, directly on point, that fuel and power for domestic use in demised areas charged on a meter direct to the tenants of flats is eligible for the reduced rate. Where a landlord owns or rents a secondary credit meter, its supplies are reduced-rated for qualifying use whether made at cost or with a mark-up. Standing charges take the rate of the supply they relate to.

Against that sits the other route. Where landlord and tenant contract for a single supply of heated accommodation, it is one supply of accommodation, taxed accordingly, exempt in the residential case, with the landlord's input tax on the fuel then generally irrecoverable. Two separate supplies are only likely where the supply is metered.

Three consequences follow. Bulk gas and electricity coming into a residential energy centre should generally bear VAT at 5% rather than 20%, because supplies to a landlord or managing agent of accommodation used for qualifying purposes are themselves reduced-rated. That holds where the energy centre serves wholly or mainly qualifying-use dwellings; mixed schemes serving commercial units alongside dwellings require apportionment, subject to the de minimis and 60% qualifying-use rules. Whether your heat income is exempt or reduced-rated depends on facts you need to establish rather than inherit. And on a portfolio, a treatment that has drifted is worth quantifying properly. In a single small block I have seen a quarter of a million pounds of gas and electricity flow through four years of accounts. Multiply that across a managed estate.

The classification carries a second consequence that is easy to miss. The Climate Change Levy exemption available for gas supplied to a qualifying use follows the same underlying facts as the VAT treatment. A misclassification that overstates a VAT liability is very often overstating a CCL liability alongside it, on the same energy volumes.

Get the classification wrong and the accounting treatment compounds it, because recoverable and irrecoverable VAT sit in completely different places in the accounts. Where input VAT is genuinely recoverable, it nets off against output VAT owed to HMRC and only the timing (when it is paid versus when it is reclaimed) touches the cash flow. Where it is irrecoverable, because the onward supply was wrongly treated as exempt, or simply is exempt, it stops being a VAT question and becomes a real cost, sitting in the profit and loss account in addition to the underlying energy spend, not instead of it. A four-year VAT misclassification is therefore not a technical filing error to tidy up later. It is a genuine cost that has been quietly overstating the deficit, or understating it, for as long as nobody has checked the heat network accounts.

When Does a Heat Network Need to Register With Ofgem?

Ofgem became the statutory heat networks regulator for Great Britain on 27 January 2026, under the Heat Networks (Market Framework) (Great Britain) Regulations 2025 (SI 2025/269), made under the Energy Act 2023 and commenced by SI 2026/7, with the authorisation conditions taking effect the same day. Every operating heat network was automatically deemed authorised on that date, so nothing stopped operating overnight.

Enforcement powers include financial penalties and consumer redress orders, carrying on a regulated activity without authorisation is a criminal offence, and operating a relevant heat network that has not been registered once the deadline below has passed puts the operator in breach of authorisation condition A4.

To be fair to Ofgem, the stated approach for now is flexible: an initial focus on working with the sector to embed the rules, with enforcement applied where proportionate, particularly where there is significant detriment to consumers. But heat network registration under authorisation condition A4 asks for information about your organisation, ownership, financial resilience and consumer and billing arrangements, and Ofgem has said it expects thousands of operators to register inside the same window.

Two conditions make the heat network accounts unavoidable rather than merely helpful. Condition A6 requires charges imposed on consumers to be fair and not disproportionate; condition A7 requires them to be structured and attributed to costs consistently with that. Demonstrating either means knowing what the network cost and what was billed. There is also a detail in the cost allocation guidance worth acting on: variable costs should be recovered through the unit rate and non-variable costs through the standing charge, and charges attributable to a regulatory penalty or consumer redress payment are presumed unfair, so those cannot simply be passed to residents.

Condition A13 requires that amounts collected from consumers for maintaining, repairing or replacing the network remain available for that purpose, and that directors have a reasonable expectation that resources will be available. A network sitting on an unquantified deficit funded by borrowing from another fund sits squarely in that territory.

Two standards worth knowing by name, because funders are already asking for them even where Ofgem is not yet enforcing them directly. The Heat Networks Code of Practice, CP1 (2020), is the technical benchmark for how a network should be designed, built and operated, and government-funded schemes are already required to confirm compliance against it as a condition of grant.

Heat Trust registration is the customer-protection equivalent, and grant conditions I have seen typically require it within one year of a network first connecting domestic or micro-business customers, with a defined alternative: an independent audit confirming an equivalent scheme, where Heat Trust registration itself is not used. Neither is currently a universal legal requirement for every network.

Both are becoming the practical standard the sector is measured against, and a network that cannot point to either is going to find that a harder conversation with funders and insurers than with Ofgem. Sitting alongside both is the Heat Network Technical Assurance Scheme (HNTAS), the statutory technical standard now being introduced, which brings its own phased metering, monitoring and certification deadlines for existing and new networks.

When is the Ofgem Registration Deadline for Heat Networks?

The registration window is longer than the headline deadline suggests. It runs from 1 April 2025 to 27 January 2027, and Ofgem's digital registration service opened in April 2026 and is now live. Every relevant heat network must complete registration by 26 January 2027. Operating without registration after that date is a criminal offence, with enforcement powers including financial penalties of up to 10% of turnover or £1 million, whichever is higher, and consumer redress orders.

So registration is the deadline everyone can see. Recoverability is the one costing money quietly. Both point at the same thing: get the Heat Network accounts done.

What Should a Well-Prepared Heat Network Account Include?

If you are assessing your own position, or someone else's work, these are the questions I would ask.

• Does the income figure tie to the billing agent's own reported total, and is the difference, if any, explained rather than plugged? 
• Can you see, per unit and per period, what was billed and on what read? Is the proportion of estimated reads disclosed?
• Are energy accruals tested, particularly where invoices were estimated and later corrected? Does the debtor reconciliation balance?
• Are the heat monies in a designated client account, reconciled?
• Is any inter-fund borrowing disclosed with the lease basis identified?
• Is the VAT treatment stated and consistent across the period?
• Where a recent year looks better than the ones before it, is that a genuine improvement in the underlying tariff and cost position, or a one-off, an insurance settlement, a change of billing agent, a catch-up demand, that will not recur?
• And for every aged deficit shown as recoverable, what was demanded, when, and were section 20B(2) notices served?

That last question is the one to ask first, because it is the only one where the answer changes depending on when you ask it.

Conclusion

Heat networks currently supply around 3% of UK heat, across more than 14,000 networks and close to half a million connections. Government policy anticipates something closer to a fifth of national heat demand by 2050, with an interim ambition of about 7% by 2035, and heat network zoning beginning to designate areas during 2026 where connection can be required.

Every one of those networks needs accounts, every year, permanently. The volume of this work is going to multiply, and the regulatory expectation for evidenced numbers is tightening at the same time. The sector is short of people who can do it properly, and I do not think that shortage resolves quickly.

For now, though, the useful action is narrower and more urgent than any of that. If you have heat networks with accounts more than 18 months behind, the priority is not registration, and it is not a system. It is finding out how much of the deficit is still legally recoverable, before the answer gets worse.

FAQs