Both London and Paris have a government-issued energy rating on every property — the UK's EPC band and France's DPE/GES score. Both governments have telegraphed the same direction of regulatory pressure (MEES landlord rules in the UK, the loi Climat phase-out of passoires thermiques in France). You might expect the two markets to price energy efficiency the same way. The data says they very much do not.
London's energy-rating factor (EPC band collapsed to a three-level scale where A/B = +1, C–F = 0, G = −1) ran up through 2022 to a peak of +3.5% — i.e. for a moment the market was paying about 3.5 percentage points more per sqm to step a property up an energy band than it had at the start of 2020. That premium has since round-tripped and then some: cumulative now −3.6%, with the trough at −8.0%. Year-on-year, the factor has returned −4.1%.
The cumulative loss since the model's start in 1995 is −9.8% — the most-negative cumulative factor return in the London model. Whatever the regulatory story, the market is, on net, not paying for it.
Paris's GES (greenhouse-gas) coefficient over the same window has barely moved. Range is −1.7% to +0.2%, current cumulative is −0.5%. The latest 12-month return is +0.3% — slightly positive but well within noise. Whatever else can be said about Paris property over the last five years, the GES factor has not been a meaningful return contributor in either direction.
| Market | Factor | Cumulative since Jan 2020 | Peak | Trough |
|---|---|---|---|---|
| London | EPC rating ({A/B, C–F, G}) | −3.6% | +3.5% | −8.0% |
| Paris | GES (greenhouse-gas) linear | −0.5% | +0.2% | −1.7% |
One plausible reading: London is a market that trades the regulatory narrative — buyers will bid up A/B-rated stock during periods of regulatory focus (the 2021–22 MEES tightening discussions are a candidate driver of the +3.5% peak), then unwind the trade when enforcement timelines slip or political signals soften. Paris, by contrast, prices the GES band largely as a fixed effect on level — better-rated flats are worth more in absolute terms, but that premium hasn't been re-rated either up or down over the cycle.
A second reading is more mechanical: French DPE methodology was substantially overhauled in 2021 and again in 2024, which may have introduced classification noise that washes out a return signal in either direction. Either way, the practical take-away for analysts comparing the two cities is that energy-rating premium is a London return factor and a Paris level factor.
Neither NYC nor Singapore has a directly comparable building-level energy-rating field, so we don't track an energy factor for them. New York's Local Law 97 produces building-level emissions data but it lives outside the DOF transaction dataset; Singapore HDB has BCA Green Mark ratings but they aren't joined to the resale flow. Adding both markets to the four-city panel for this analysis is on the roadmap.
Energy-rating coefficients, p-values, and monthly returns.
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; cumulative return is the geometric change since January 2020. See About for the full spec.