England & Wales · Full EPC register · Research

Rental vs owner: the EPC gap, on the full register

The familiar claim is that private-rented homes are the worst for energy efficiency. Across all 16,718,861 rated dwellings in England & Wales, the raw numbers say the opposite — and once you hold construction age constant, the picture is clearer still.

Free to cite with attribution · Open Government Licence.

60.5%

Owner-occupied below band C (8,462,639 homes)

53.0%

Private-rented below band C (3,359,528 homes)

32.7%

Social-rented below band C (2,805,143 homes)

current stock (latest certificate per UPRN), England & Wales domestic. Tenure as recorded on the EPC. Certificates 2014-2026. Source: EPC Open Data (Open Government Licence).

Cite this

On the full EPC register for England & Wales, 60.5% of owner-occupied homes are below EPC band C, against 53.0% of private-rented homes. The Home Index, 2014-2026. Source: EPC Open Data (Open Government Licence).

Free to cite with attribution · Open Government Licence.

The headline

Owner-occupied homes are further from band C, not rented ones

Owner-occupied stock has a 7.5-point larger below-C share than private-rented, and social housing is best of all. That is a composition effect: rented stock skews to newer, smaller flats, social landlords have retrofitted at scale, and the Minimum Energy Efficiency Standard has lifted the worst rental stock since 2018.

The controlled comparison

Age-for-age, rented homes are the better ones

The honest test holds construction age constant. In 10 of the 12 age bands, private-rented homes have a lower below-C share than owner-occupied — the rental penalty all but disappears, and reverses, once you compare like with like. Only the newest new-builds (where owners dominate) tilt the other way.

Construction ageOwnerPrivate rentedRented − owner
Pre-190086.0%76.7%-9.3 pp
1900–192979.9%70.3%-9.6 pp
1930–194977.7%67.2%-10.5 pp
1950–196669.4%59.4%-10.0 pp
1967–197567.1%56.7%-10.4 pp
1976–198253.3%43.4%-9.9 pp
1983–199050.2%40.7%-9.5 pp
1991–199550.7%39.6%-11.1 pp
1996–200236.6%23.9%-12.7 pp
2003–200610.6%10.0%-0.6 pp
2007–20116.1%7.0%+0.9 pp
2012+3.3%8.6%+5.3 pp

Below-C share within each tenure × construction-age cohort (cohorts with at least 300 dwellings). A negative final column means private-rented homes of that age are more efficient than owner-occupied ones.


For local and regional news desks: click any area for its full report and citable figure.

The local picture

…but in some areas the rental gap is real

The national average hides wide local variation. These are the authorities where private-rented stock is furthest behind owner-occupied — the places where the 2030 MEES band-C rule will bite landlords hardest.

Widest rental gap (PRS minus owner, below-C share)
#Local authorityGap
1Uttlesford
+10.4 pp
2Rutland
+8.9 pp
3Mid Suffolk
+7.7 pp
4Derbyshire Dales
+7.1 pp
5Melton
+6.5 pp
6Cotswold
+5.6 pp
7Ribble Valley
+5.5 pp
8South Derbyshire
+4.5 pp
9Vale of White Horse
+4.0 pp
10City of London
+3.9 pp
11Tandridge
+3.9 pp
12South Norfolk
+3.3 pp
…and where rented stock is ahead of owner-occupied
#Local authorityGap
1Rushmoor
-22.6 pp
2Epsom and Ewell
-17.1 pp
3Bromley
-16.9 pp
4Reigate and Banstead
-16.0 pp
5Warrington
-15.5 pp
6Three Rivers
-15.2 pp
7Welwyn Hatfield
-14.5 pp
8Gateshead
-14.4 pp
9Surrey Heath
-14.3 pp
10Havering
-14.2 pp

What's at stake

Fusing the register with HM Land Registry sales, the below-C housing stock of England & Wales carries an estimated £1.7tn in matched-sale value — a modelled estimate. For a landlord, the question isn't the national average but their own properties.

Method & limits

EPC tenure is self-declared at assessment; an area's tenure mix in the register need not match its census mix, so cuts describe rated stock, not a census. rental penalty = private-rented gap minus owner-occupied gap (percentage points). Each dwelling is counted once at its latest certificate. Tenure is self-declared at assessment and an area's register mix need not match its census mix, so these are cuts of rated stock, not a tenure census. Areas with fewer than 300 owner or private-rented dwellings are not ranked.