The expected pattern
The intuition most buyers carry is: wealthy areas have efficient homes, deprived areas have draughty ones. The intuition is half-right.
If you take every postcode in England with at least five EPC certificates lodged since 2018, join it to the LSOA-level Index of Multiple Deprivation 2025, and bucket by decile, the picture is much flatter than the intuition suggests:
IMD decileExpected mean EPC (current efficiency, /100)
3 (more deprived)64.8
465.5
564.6
665.2
765.4
866.0
966.2
(Source: ONS 2% systematic EPC sample joined to MHCLG IMD 2025 at LAD level. Deciles 1, 2 and 10 are noisier and excluded from the baseline because the sample is dominated either by individual buildings of unusual character — high-rise social blocks at one end, large heritage homes at the other.)
The gap between decile 3 and decile 9 is 1.4 EPC points — roughly the difference between a band-D house with single glazing and the same house with double glazing on the windows. That's much less than people assume. The reason is straightforward: most English housing stock pre-dates the era of strong energy regulation, regardless of who lives in it. Wealth doesn't buy a new build; it buys a refurbished Edwardian terrace, which usually still has a D rating.
But that flat national average hides outliers. Our sample of ~15,500 postcodes contains roughly 5% that break the pattern by ≥15 EPC points in either direction. They are the most interesting postcodes in the country, for two opposite reasons.
Methodology
+Per postcode, we computed the mean current-energy-efficiency score across all EPCs lodged since 1 January 2018 (excludes ageing pre-MEES certificates).
+We required ≥5 EPCs to register in the dataset; ≥8–10 to qualify as an outlier candidate.
+We joined each postcode to its LSOA via ONS postcode directory November 2025, then to IMD 2025 decile via MHCLG LSOA file (33,755 LSOAs).
+We compared each postcode's mean EPC to the national expected value for its IMD decile (per the table above).
+Positive outlier = postcode in decile 1–4 (more deprived) with mean EPC ≥15 points above expected.
+Negative outlier = postcode in decile 7–10 (less deprived) with mean EPC ≥15 points below expected.
This is England only. The sample so far covers ~15,500 postcodes spanning London, Liverpool, Manchester, Birmingham, the North East, parts of the East Midlands and Yorkshire. National rollout is ongoing.
Caveats:
+"EPC current efficiency" is a model output, not a measured kWh figure. Inputs can be wrong — wall-construction U-values are particularly prone to error.
+A high-EPC postcode usually means it's dominated by recent new-builds. That's not the same as saying every property in the postcode is efficient — period properties in the same postcode will be much lower.
+IMD operates at LSOA scale (about 1,500 people). A postcode in a deprived LSOA can still be a high-end gated estate.
Top 25 "punching above their weight" (deprived but efficient)
The pattern that jumps out: large modern flat developments dropped into otherwise-deprived LSOAs.
PostcodeLADIMD decileMean EPCΔ vs expectedEPCs
L11 9EDLiverpool195.4+27.937
L11 9EELiverpool195.3+27.825
L11 9DXLiverpool192.1+24.640
M13 9PJManchester489.9+24.414
M16 9GHTrafford291.2+24.213
NW2 2FRBarnet388.9+24.112
NW4 1BEBarnet489.2+23.719
NW1 8QJCamden290.1+23.120
NW10 6SGHammersmith and Fulham387.5+22.7126
NW6 4BRCamden289.1+22.160
M12 5XQManchester189.3+21.855
NW1 9SYCamden386.3+21.523
NW2 2ERBarnet386.3+21.520
M11 3DFManchester288.5+21.515
NW1 4BHCamden386.0+21.225
NW1 0NZCamden386.0+21.224
NW10 6FJHammersmith and Fulham385.9+21.160
M13 9JYManchester486.6+21.112
M13 9SGManchester188.5+21.017
NW10 6FNHammersmith and Fulham385.8+21.0178
NW10 7GTBrent486.5+21.065
NW10 7GXBrent486.5+21.042
NW1 9GPCamden385.8+21.024
M12 5FYManchester188.4+20.916
NW10 7GZBrent486.4+20.990
These are almost entirely new-build flat blocks. NW10 6FN (the largest in our sample with 178 EPCs) is Imperial Wharf-adjacent in Fulham — a 2010s riverside scheme dropped into a Hammersmith LSOA whose deprivation score reflects nearby social-housing estates. M11/M12/M13 Manchester are new student-accommodation and city-centre living developments. L11 in Liverpool is the same — a single regeneration scheme sitting in an otherwise-deprived LSOA.
The signal for buyers: a postcode that looks "wrong" for its IMD decile usually contains a recent development. That can be a fantastic energy-efficient buy at a sub-postcode-average price. It can also be a service-charge nightmare. Pull the leasehold details before celebrating.
Top 25 "punching below their weight" (affluent but inefficient)
The opposite pattern: period housing in expensive postcodes, where retrofit hasn't happened.
PostcodeLADIMD decileMean EPCΔ vs expectedEPCs
L18 3HZLiverpool946.6-19.68
NW3 5NNCamden (Hampstead)1046.1-19.48
L15 6TSLiverpool847.5-18.58
EC4A 1BYCity of London847.9-18.136
L14 6TTKnowsley747.7-17.716
NW3 7XFCamden949.2-17.010
NW11 6EDBarnet (Golders Green)748.6-16.811
L16 1JBLiverpool849.4-16.610
NW1 8UYCamden849.5-16.512
NW6 1EACamden850.2-15.810
B13 8LPBirmingham (Moseley)850.4-15.69
L16 8NWLiverpool951.5-14.711
NW3 2SJCamden (Hampstead)951.7-14.59
L16 7PPLiverpool951.9-14.38
L14 0JLKnowsley751.2-14.29
N6 6LRCamden (Highgate)751.6-13.814
NW11 6HGBarnet (Golders Green)852.2-13.812
EC1Y 0SNCity of London852.3-13.722
L14 7PTKnowsley852.3-13.79
NW3 7NTCamden952.6-13.68
B17 9QXBirmingham (Harborne)953.0-13.211
L18 3EBLiverpool1052.5-13.08
NW3 7ESCamden853.0-13.015
N6 6JSCamden (Highgate)752.5-12.98
L16 5EBLiverpool853.1-12.99
NW3 (Hampstead) and N6 (Highgate) dominate the affluent-but-inefficient list. The reason is simple: large Victorian and Edwardian detached houses on solid-brick walls, with conservation-area restrictions on visible insulation, sash windows that can't easily be replaced with sealed double-glazed units, and no incentive to retrofit because the gain doesn't show up in the asking price.
The fascinating thread running through both ends of the table is that Camden appears 11 times across the two lists, with five entries in each direction. The same borough that contains some of the most efficient postcodes in the country (NW1 modern blocks, NW6 conversion estates) also contains the most stubbornly inefficient (NW3 Hampstead villas). EPC and IMD decile pull in opposite directions at the borough level — the wealth is in the period stock, but the retrofit opportunity sits there too.
Policy implications
A few things worth saying out loud, given the Government's MEES target (all rented homes EPC C by 2030):
1.The retrofit opportunity is in the affluent decile. The bottom-25 list above is the actual policy challenge — well-resourced households living in 1900-era stock that should be at EPC C and is at D-, paying disproportionate gas bills.
2.Targeting subsidy at deprivation decile is the wrong cut. A scheme that targets IMD decile 1–3 will largely subsidise *new-build flat owners*, not deprived households living in old stock. The targeted homes are physically present but they're under-represented in the postcode-mean — because the new-build mean dominates.
3.Conservation areas should be allowed sealed sash-style double-glazed windows. A meaningful fraction of the bottom-25 postcodes sit inside conservation areas where window replacement requires consent. Where consent is refused on aesthetic grounds, the retrofit gap widens permanently. (Heritage bodies disagree; we're flagging the trade-off, not resolving it.)
4.The new-build EPC ≠ delivered performance. Recent academic work (UCL Energy Institute, 2024) suggests post-2010 dwellings perform 15–25% worse than their EPC claims. That doesn't change the *relative* ranking we've shown but it does mean the "+25 vs expected" headline figures should be read as "+15" in measured-energy terms.
What this means for buyers
If you're shopping in Hampstead, Highgate, Golders Green, Harborne, or comparable affluent period suburbs:
+Assume the EPC will be 45–55 unless the listing photos show triple-glazed sashes and a visible heat pump.
+Budget £30–80k for a serious retrofit (wall insulation, glazing, heating system, ventilation).
+Check whether the property sits in a conservation area; if it does, factor a 6-month consent process and possible heritage-officer pushback.
If you're shopping in NW10, M11, L11 or similar regeneration-zone postcodes:
+The high EPC is real and currently unpriced — well-built modern flats often sell ~5–8% below the comparable unit in a "better" postcode despite running on 30–40% less energy.
+Check the cladding situation. Many of these blocks are post-2010 high-rises with ACM/HPL questions that haven't fully resolved.
+Service charges in these schemes can run £4–8k/year. Pull the management pack before offering.
Outcode-level dossiers for the postcodes mentioned above:
Sources
+EPC certificates: MHCLG Open Data, all lodgements ≥ 1 January 2018
+Postcode to LSOA: ONS Postcode Directory November 2025
+National EPC-by-IMD baseline: 2% systematic sample, EPC × IMD 2025 join, computed for this analysis
Related