← Back to blog

Energy rating and property value

Does a better EPC band actually translate into a higher per-sqm price — and by how much, in which markets, over what window? CityDataLab estimates a rolling cross-sectional regression on hundreds of thousands of transactions per month, so we can isolate the energy-rating premium from everything else that's moving in the housing market. This post collects what the data says, across the cities where energy ratings are recorded.

The energy premium right now

The table below shows the latest read on each city's energy-rating coefficient. The "cumulative" column is the return since the model's first month (1995 for London, 2017 for Paris); "12-month return" is computed geometrically on the cumulative curve, not a difference of cumulative percentages.

CityFactorCumulative12-month return
London EPC rating (A/B = +1, C–F = 0, G = −1) -9.8% -4.1%
Paris GES (greenhouse-gas) score, linear encoding +0.7% +0.3%

How the two markets have priced energy efficiency since 2020

Cumulative return on each city's energy-rating coefficient since January 2020. Latest London: Oct 2025. Latest Paris: Apr 2025.

London bid the EPC premium up through 2022 (peaking at +3.5% above its Jan-2020 baseline), then round-tripped and continued through to a deep negative. Paris's GES coefficient over the same period has been close to flat. The same regulatory direction — MEES in the UK, the loi Climat passoires thermiques phase-out in France — has produced opposite market reactions.

Why this data is hard to find anywhere else

Headline house price indices mix two different things: how much actual buyer demand has shifted, and how the mix of properties getting sold has rotated. To isolate the energy-rating premium specifically, you need a hedonic regression that controls for floor area, room count, construction era, location and tenure all at once. We run that regression every month for every city we cover, and we publish the resulting time series of factor returns — including the energy-rating coefficient — back to the start of each city's dataset.

For London that means 30 years of data, joined from HM Land Registry transactions to MHCLG EPC certificates by postcode. For Paris it's eight years of DVF transactions joined to government-issued DPE/GES certificates.

Related reading

See the energy-rating factor in each city's full model

Track every factor, every month

Energy rating, size, freehold/leasehold, construction era — with rolling p-values and 30 years of data.

Open the global comparison →

Methodology: rolling 3-month cross-sectional OLS of log(price/sqm) on hedonic property characteristics per city. The energy-rating coefficient is the per-band premium on each city's encoding (London EPC collapsed to A/B = +1, C–F = 0, G = −1; Paris GES on the published linear scale). See methodology & data sources for the full spec.