We Are Building the Next Retrofit Programme

What is already true

The Climate Change Committee commissioned Arup to assess overheating across the UK housing stock. The resulting report, Addressing overheating risk in existing UK homes (October 2022), found that under current weather conditions around 55% of UK homes — roughly 15.7 million — already fail the bedroom overheating criterion in CIBSE TM59. That criterion is not dramatic: a bedroom's operative temperature should not exceed 26°C for more than 1% of annual night-time hours, measured between 10pm and 7am.

Fifteen point seven million homes fail that today. Not by 2050. Today.

The Committee's more recent assessment, A Well-Adapted UK (20 May 2026), projects that 92% of existing UK homes are likely to overheat by 2050 at around 2°C of global warming. It puts heat-related deaths on a path from the low thousands a year today to between 3,000 and 10,000 by mid-century without adaptation.

And this is not a projection about a distant climate. By 15 July this year the UK had recorded more days above 30°C than in the whole of 1976, and 2026 became the first year on record to reach 35°C in May, June and July. Overnight minimum temperatures are tracking at a record high for this stage of the summer, on a series going back to 1931.

Overnight minima are the number that matters here. A house that cannot shed its heat after dark never resets. That mechanism — and why modern, well-insulated homes are the worst affected — is the subject of our earlier piece on why UK houses get so hot in summer. This article is about something narrower and, I think, more uncomfortable: what we are still building while we know all of the above.

The homes going up this year are not the solution

It is natural to assume that whatever is wrong with the existing stock, at least the new estates going up around every market town are being built properly. That assumption does not survive contact with what the regulations actually require.

A new home in England can be built entirely legally to the limiting values in Table 4.1 of Approved Document L, Volume 1:

It is worth pausing on that window line, because it is the one most people can picture. A backstop of 1.6 W/m²K is a double-glazing number. A good triple-glazed unit of the kind treated as ordinary in German, Austrian or Polish practice sits around 0.8 W/m²K — roughly half the heat loss, and, more importantly for this article, a unit whose solar heat gain coefficient can be specified by orientation. The typical volume housebuilder window in the UK remains a PVCu double-glazed casement.

These are the backstops — the worst a new home can legally be. A house built to them is not defective and not mis-sold. It is compliant. It is also, on the fabric numbers, roughly what a competent builder in several continental European markets would treat as a starting point rather than a ceiling. That gap is the whole reason APMBuild sources through a European supply chain rather than a UK merchant catalogue.

Are new homes in England assessed for overheating risk?

Some are, and a great deal of building work is not. The short answer is that overheating risk in new homes in England has been a regulated question only since 15 June 2022, that the requirement covers new residential buildings and nothing else, and that the assessment can be satisfied by a simplified method which never models your actual building. The detail matters, so here it is.

Approved Document O has only applied since 15 June 2022, and its own Application section (paragraph 0.3) is unambiguous about how far it reaches:

"The guidance in this approved document applies to new residential buildings only."

Table 0.1 of that document lists what counts: dwellinghouses and flats, residential institutions, and student accommodation — all of them new. Work to an existing home is a different matter. An extension or a conservatory bolted onto a house built in 1994 is not a new residential building, and the summer performance of the glazing you are adding is not assessed under Part O at all. That is precisely where the most aggressive glazing goes in — the full-width bi-fold, the lantern rooflight, the south-facing garden room. We build a lot of those, which is why every APMBuild extension gets the summer question asked whether or not a regulation demands it.

And everything completed before 15 June 2022 — including the whole post-2010 wave of airtight, well-insulated, heavily glazed housing — was built with no summer performance requirement of any kind. Nobody modelled it, because nobody had to.

And the system we are building it out of

There is a second variable, and it is the one I find hardest to raise on a site without an argument starting.

Part L tells you what the envelope must achieve. It says nothing about what the envelope is made of. And the single property that does most to flatten a summer temperature swing — thermal mass — appears nowhere in a Part L compliance calculation, because Part L is a heat-loss calculation and mass does not reduce annual heat loss.

The differences between systems are not marginal. Using the figures set out in our own lifecycle comparison, Timber Frame vs Masonry, expressed per m² of floor area:

A heavy building changes temperature slowly. It absorbs daytime heat into the fabric and gives it back overnight, when the outside air is cooler and the windows can safely be opened. A lightweight building tracks the air temperature inside it almost directly: minutes to heat up, and then a long evening of waiting.

Here is where this section joins the first one. The whole night-purge strategy — the thing Approved Document O's simplified method leans on, the thing every "just open the windows" answer assumes — depends on the outside air being meaningfully cooler than the inside air after dark. The Met Office series this summer has overnight minimum temperatures at a record high for the stage of season, on data going back to 1931. A lightweight house with no mass, in a run of warm nights, has nothing left to fall back on. It did not store the coolness of the previous night, because it cannot.

And the exposure is growing rather than shrinking. The Structural Timber Association reported in November 2025 that timber frame's share of new build in England had risen from around 12% to 15% in a year, and that in Scotland timber frame is already close to saturation as the dominant system. Industry forecasts model a further rise toward 25% by 2030. Whatever else that shift achieves — and it achieves real things on programme and on at-gate embodied carbon — it removes thermal mass from an increasing share of the national housing stock at exactly the moment the climate argument for keeping it is strengthening.

I want to be careful here, because it would be easy to read this as a pitch for masonry. It is not, and our own lifecycle piece is explicit that timber used well — cut-rafter roofs, cladding from local sources, wood-fibre insulation — is among the best material choices available. A timber-frame house with proper external shading, disciplined glazing by orientation and a working purge route will outperform a masonry house that has none of those things. Mass is a buffer, not an absolution.

The objection is narrower and it is the same objection as the rest of this article. The structural decision is taken on programme, cost and at-gate carbon. The summer consequence is not priced into it, because no calculation in the compliance process asks the question. So the buffer quietly disappears, nobody records that it has, and the cost of its absence lands on a household in about fifteen years' time.

Why the market produces this outcome

This is the part that is usually described as cynicism and is in fact documented.

In its final report on the housebuilding market in Great Britain (26 February 2024), the Competition and Markets Authority concluded that the market is not delivering well for consumers on either the number or the quality of homes; that housebuilders lack strong incentives to compete on quality; that snagging reports from owners have risen over the past decade; and that buyers frequently struggle to obtain accurate information about what they are purchasing.

That is the regulator, not a competitor's opinion. And it is worth noting what the industry's own quality measure does and does not capture. The Home Builders Federation's National New Homes Customer Satisfaction Survey asks buyers about their home roughly eight weeks after they move in. A buyer who completes in November is answering about a house they have never experienced in August. Whatever that survey measures, it cannot measure summer performance — and no industry metric currently does.

The mechanism is not villainy. It is that the person who pays for a better building envelope is not the person who benefits from it. A developer's return is realised at sale. Thermal comfort in August 2041 does not appear anywhere in that calculation, and no buyer has ever paid a premium for a psi value they were never shown.

It is worth being precise about where architects sit in this. On a volume housing site the architect is frequently out of the process after planning consent, working to a standard house type, or holds a design-and-build appointment under which specification decisions pass to the contractor. Value engineering happens downstream of design far more often than within it. Most architects I deal with are arguing for the shading, the glazing spec and the summer modelling — and losing that argument to a budget line. The problem is structural, and design professionals are mostly on the same side of it as the client. That is why our work with architects and developers starts at concept stage rather than at tender.

The buyer cannot see any of this

Here is what a purchaser can assess on a show-home visit: the kitchen, the finish, the room sizes, the garden, the road.

Here is what they cannot assess, and are generally not shown:

• whether the dwelling was assessed against Part O by the simplified method or by dynamic thermal modelling — and if the simplified method, that no external shading, ventilation strategy or building-specific geometry entered the calculation at all
• the modelled summer performance of the bedrooms
• the as-built air permeability result rather than the design target
• how thermal bridges were detailed at junctions, and whether accredited construction details or conservative default psi values were used
• whether the purge ventilation strategy still works once ground-floor openings are discounted for security, as Approved Document O Section 3 requires

An EPC does not report any of it. An EPC is a heating-season metric. It has nothing to say about whether a bedroom sits at 27°C at two in the morning in August. That limitation survives the coming methodology change, too: as we set out in the Home Energy Model guide, HEM raises the resolution of the energy calculation enormously, but the compliance question it answers is still fundamentally about energy and carbon, not about whether you can sleep.

So the honest description of a new-build purchase, on this specific issue, is that the buyer is making a six-figure decision on information they have not been given and would struggle to obtain. That is not their failure. It is a disclosure gap.

What it costs to fix later

The CCC/Arup work also priced the retrofit. Mitigation measures for overheating in existing homes were costed at upwards of £10,000 per dwelling (2022 prices), and the report was explicit that the scale of works implied across the stock cannot be delivered to all dwellings at once — construction industry capacity simply does not allow it.

Two consequences follow, and they are the whole argument of this piece.

First, there will be a queue, and homes built today will be in it. The CCC's A Well-Adapted UK puts the required adaptation investment at around £11 billion a year to the 2050s, split broadly evenly between public and private funding.

Second, the cheap fix has an expiry date. Arup found that at 2°C of warming, passive measures alone — shading, ventilation, glazing changes — can address overheating in up to 80% of dwellings. At 4°C, passive measures run into a hard limit and mechanical cooling becomes unavoidable for a much larger share of the stock. Air conditioning in a poorly shaded, lightweight, airtight house is an operating cost for the rest of that building's life, plus a load on a grid that is simultaneously being asked to absorb heat pumps and electric vehicles.

Compare that with the cost of the same decisions taken at concept stage: glazing area and orientation, an eaves overhang or external shutter, a purge route that survives the security discount, a ventilation strategy sized properly. At RIBA Plan of Work Stage 2 most of that is drawing-board work. After planning consent it is expensive. After handover it is a scaffold, a planning application and a disrupted household — as our deep retrofit cost breakdown sets out element by element.

There is a second-order point here too. Lightweight construction is not neutral in this equation: thermal mass is one of the few things that flattens the daily temperature swing without any moving parts, and the specification choice that removes it is usually made on embodied-carbon or programme grounds without the summer consequence ever being priced. We looked at that trade-off in full in Timber Frame vs Masonry: The Full Lifecycle View, which examines a similar gap between what a specification is marketed as and what it does over sixty years.

This is the same shape as the insulation story, thirty years earlier in its cycle. We know how that one ended, because we are still paying for it.

What to ask before you sign

If you are commissioning a house, buying a new-build, or advising someone who is, these questions cost nothing and are answerable by any developer or contractor who has done the work:

1. Was Part O demonstrated by the simplified method or by dynamic thermal modelling? Can I see the output?
2. What is the modelled peak bedroom temperature in summer, and against which weather file?
3. What is the as-built air permeability test result for this plot — not the design target?
4. What external shading is specified, and was it modelled or just drawn?
5. Which openings were counted as securely openable for night-time ventilation?
6. Were accredited construction details or default psi values used for thermal bridging?

If the answers do not exist, that is itself the answer, and it is worth knowing before exchange rather than during the third heatwave of the summer.

None of this requires anyone to build to Passive House standard. It requires the summer question to be asked at the point where it is still cheap to answer. On our own projects that is what PHPP modelling is for — it returns an overheating frequency figure at design stage, when the fix is still a line on a drawing rather than a scaffold.

If you are buying rather than building, our homeowner guidance and calculators are a reasonable place to start. If nobody on your project team can answer the six questions above with numbers, that is the work we can help with. Email office@apmbuild.uk or call 07711 266 107.