London's new-build flats have a well-documented problem: across all 33 boroughs, a flat with a recorded new-build sale is far less likely to show up with an Energy Performance Certificate than an older resale flat (London New-Build Flats Have Lower EPC Coverage in Every Borough). A fresh pass through the same data for houses finds the opposite: new-build houses carry slightly more EPC coverage than resale houses, not less, and the gap that defines the flat market barely exists for detached, semi-detached and terraced homes.
The headline numbers
Across London's 1,068,843 house addresses (detached, semi-detached and terraced, drawn from HM Land Registry Price Paid Data and matched against the EPC register, fetched 5 November 2026), coverage splits as follows:
| Property type | Resale coverage | New-build coverage | Ratio |
|---|---|---|---|
| Flats (previously published) | 29.7% | 9.1% | 3.25x lower |
| Houses (this analysis) | 70.4% | 74.0% | 1.05x higher |
A house with at least one new-build sale on record is matched to a certificate 74.0% of the time (34,839 of 47,053 addresses), against 70.4% for houses with no new-build sale in the data (719,349 of 1,021,790). That is a 3.6-percentage-point gap in the opposite direction to flats, where new-build coverage trails resale coverage by 20.6 points — roughly a sixth the size, and running the other way.
The same borough, two different stories
The clearest way to see this is to compare a borough's flat gap with its house gap side by side. The four boroughs with London's widest new-build/resale flat gaps show almost no equivalent gap on the house side:
| Borough | Flat coverage gap (resale ÷ new-build) | House coverage gap (new-build ÷ resale) |
|---|---|---|
| Harrow | 8.9x | 1.13x |
| Havering | 8.2x | 1.08x |
| Bromley | 7.2x | 1.06x |
| Hillingdon | 6.0x | 1.11x |
In each case, the borough's new-build flats are many times less likely to carry a certificate than its resale flats — but its new-build houses are, if anything, marginally more likely to. Across all 32 boroughs with a usable house sample, 23 show new-build houses with higher coverage than resale houses and 9 show the reverse, but none of the 9 comes close to a flat-scale gap: the widest house-side shortfall is Camden at 0.93x (54.6% new-build vs 58.5% resale, n=238 new-build), and the strongest new-build advantage is Hammersmith and Fulham at 1.20x (83.8% vs 70.0%, n=290 new-build). Other large, well-sampled boroughs sit close to parity either way — Croydon (1.13x, n=1,699 new-build), Bexley (1.09x, n=2,587) and Barking and Dagenham (1.05x, n=2,004) all lean mildly toward higher new-build coverage, while Lewisham (0.94x, n=1,064) and Kingston upon Thames (0.94x, n=1,636) lean mildly the other way. None of these approaches the 6x-9x gaps found on the flat side of the same boroughs. The City of London is excluded from any ranking here — its house sample (81 resale, 6 new-build) is too thin to support a comparison.
Why this probably isn't about compliance
An EPC is generated at the point a home is built, sold or let, not on a fixed renewal schedule (Energy Performance of Buildings (England and Wales) Regulations 2012), so a missing certificate in this data means no certificate has been matched to that address — not that the home is non-compliant or inefficient to live in. The flat-side gap, confirmed across every London borough, has previously been linked to how large new-build developments are registered: point blocks and towers often carry many similarly-formatted addresses lodged in a short window, which appears to make automated matching between the Land Registry record and the EPC register harder than it is for a standalone or small-terrace house. New-build houses are rarely built at that density or registered in that pattern, which is consistent with why the matching gap that dogs flats largely doesn't carry over to them.
What this means if you're checking a specific address
If a new-build flat you're looking at shows no energy rating on Homecost, that absence is common and does not by itself say anything about the property's condition — check the EPC register directly or ask the seller or agent for the certificate reference. New-build houses are much less likely to run into this gap in the first place, which is itself useful to know before assuming a missing rating is unusual or a red flag.
A real example: HA7 1DA in Harrow has 61 recorded house sales, mostly terraced homes, with ratings on file for the majority of addresses — typical of the high house-coverage, low-flat-coverage split found across the borough as a whole.
This finding builds on the borough-by-borough work already published on London's flat-versus-house EPC coverage split, which first identified how widely flat and house coverage can diverge within the same local authority. For more on how EPC data behaves across property types and cities, browse the cost intelligence section of the Homecost blog.