In 2024, a house rated G sold on average 25% cheaper than a comparable house rated D; for a flat, the gap reached 12% (the "green value" — valeur verte — study by Notaires de France, 2024 transactions). This is no longer a market nuance: it is a structural discount, one that has widened every year since 2021 and that notaries now measure on real transaction after real transaction.
If you own a thermal sieve (passoire thermique) — a home rated F or G on the DPE (Diagnostic de Performance Énergétique — the French Energy Performance Certificate, EPC) — the question is no longer whether the market discounts it, but by how much, and whether the gap will widen further as the rental bans of 2028 and 2034 approach. The latest edition of the Notaires' study gives solid figures; the edition covering 2025 transactions, still awaited, will confirm the trajectory. Here is what the figures say, and how to read the trend without over-interpreting it.
House rated G: -25% vs D. Flat rated G: -12% vs D (Notaires de France, 2024 transactions). Conversely, a house rated A sold on average +17% above an equivalent D.
What this article covers
What the Notaires de France "green value" study measures exactly and its method, the discount by class for houses and flats (F/G and A/B compared with D), the marked regional gaps, the trend since 2021, the regulatory framework feeding the trend (rental bans 2025-2034, the EU EPBD directive, the 2026 reform of the electricity coefficient), a worked example, the interpretation traps and a forward-looking — conditional — reading of the coming 2026 edition.
The Notaires de France "green value" study: what it really measures
"Green value" (valeur verte) is the share of a property's value attributable to its energy performance, all else being equal. Notaires de France have measured it for several years from their real-transaction databases (the authenticated deeds aggregated in the BIEN and PERVAL databases), which makes it the most robust reference source on the subject: it rests on actual sale prices, not on estimates or listings.
The method isolates the effect of the energy label on price through an econometric model that neutralises the other characteristics (surface area, location, property type, construction era). The result is expressed as a price gap relative to a reference class — class D, the median of the housing stock. It is this neutralisation of other factors that allows talk of a "clean" effect of energy performance, rather than a mere correlation.
The latest available edition covers the 2024 transaction year and was released by Notaires de France in late 2025. A separate market note confirmed in early 2025 that the share of sales of very inefficient homes (F-G) had stabilised at around 15% of existing-home transactions, after a peak of 17% in 2023 — a sign that the sieve market has begun to tighten.
| Indicator | Figure | Source |
|---|---|---|
| Share of existing-home sales rated F or G | ~15% in 2024 (after 17% in 2023, 11% in 2021) | Notaires de France, market note |
| Share of E-F-G (inefficient) sales | ~40% of existing-home transactions | Notaires de France, 2024 transactions |
| Model reference class | Class D (median of the stock) | Notaires de France, green value method |
Important: the study measures an average gap at national or regional scale. It does not say that your property will suffer exactly this discount — green value is a statistic, not an individual valuation. For your home, only a calculation factoring in its local market price, its class and its location gives a usable order of magnitude.
Discount by class: houses and flats do not react the same way
The first lesson of the study is that houses are far more penalised than flats by a poor label. The reason is both physical and economic: a detached house exposes more heat-losing surfaces (walls, roof, floors) and its buyer bears the renovation bill alone, whereas a flat shares part of the works at the level of the homeowners' association (copropriété).
The key figures from the 2024 transactions
On 2024 sales, relative to an equivalent property rated D, Notaires de France record the following gaps. For houses: a class A sold around +17%, while a class G suffered an average discount of -25%. For flats, the effect is more contained: a G's discount stood at around -12% versus a D. Aggregating the extremes, a high-performance house (A/B) sold for around 1.9 times the price of a sieve (F/G), against a ratio of about 1.4 for flats.
| DPE class (vs D) | Houses | Flats |
|---|---|---|
| A (very high-performance) | ~ +17% | Clear premium, more moderate |
| E | ~ -9% (2025 figure) | ~ -4% (2025 figure) |
| G (sieve) | ~ -25% | ~ -12% |
| Gap A/B ↔ F/G (2023 data) | ~ 39% | ~ 25% |
Profile: owner of a country house rated G in a slack-market region — the most exposed case, where the discount often exceeds the national average and where the buyer prices in the cost of a major renovation from the outset.
The "class E" effect, an early signal of worsening
The most telling indicator for the trend is not class G — already heavily discounted — but class E. On the most recent data, a house rated E sold around 9% cheaper than a comparable D house, against only -5% on 2021 transactions. For flats, the gap moved from around -2% to -4% over the same period. In other words, the discount is gradually "contaminating" classes that used to be near-neutral: this signals a diffuse worsening, not just a widening at the G extreme.
Regional gaps that change everything
The national average masks considerable gaps. On 2023 data, the price gap between a high-performance house and a sieve reached as much as +51% in Nouvelle-Aquitaine, against only +18% in Île-de-France. The logic is constant: the tighter the market (strong demand, scarce supply), the more the sieve discount is diluted by price pressure; the more slack and rural the market, the more the label weighs. In the Paris inner suburbs, F/G flats sold with a discount limited to a few points, whereas the most efficient homes showed a slight premium.
Why the discount should keep worsening
The worsening observed is no market accident: it is fed by a regulatory calendar that mechanically restricts the possible uses of a sieve. Three forces combine.
The 2025-2034 rental-ban calendar
The Climate & Resilience Act of 22 August 2021 (Law no. 2021-1104, art. 160) wrote into the French Construction and Housing Code (CCH) a progressive calendar banning the rental of energy-indecent homes. Homes rated G can no longer be offered for rent since 1 January 2025; those rated F will follow on 1 January 2028; those rated E on 1 January 2034. Each deadline strips rental use value from a slice of the stock, and therefore market value: an investor no longer buys a sieve at the same price knowing they will soon be unable to rent it without works.
| Deadline | Class banned from rental | Reference |
|---|---|---|
| 1 January 2025 | Class G | Law no. 2021-1104, art. 160; CCH |
| 1 January 2028 | Class F | Law no. 2021-1104, art. 160; CCH |
| 1 January 2034 | Class E | Law no. 2021-1104, art. 160; CCH |
European pressure: the 2024 EPBD directive
Above French law, Directive (EU) 2024/1275 of 24 April 2024 on the energy performance of buildings (the recast EPBD) sets a course for full decarbonisation of the building stock by 2050, with a reduction in the average primary-energy consumption of homes of 16% by 2030 and 20% to 22% by 2035. Its transposition by member states is due by 29 May 2026. This European framework locks in the trajectory: there will be no reversal on the performance requirement, which durably anchors the premium for virtuous homes and the discount on sieves.
The 2026 side effect: the electricity coefficient reform
One technical reform, however, blurs the reading. On 1 January 2026, the primary-energy conversion coefficient for electricity (the 2.3 factor applied to electricity consumption in the 3CL-DPE method) was reduced to 1.9 by the ministerial order of 13 August 2025. As a result, many electrically heated homes gained one to two classes with no works at all, sometimes leaving the sieve category. This changes nothing about the real quality of the building, but it shifts the statistical perimeter of sieves — a point to factor in before interpreting the 2026 green value edition.
⚠️ Warning: before estimating your property's discount, check its DPE class applicable in 2026. An electrically heated home rated F in 2024 may have moved up to E (or even D) on 1 January 2026 by the sole effect of the coefficient reform. The "sieve" discount then no longer applies in the same way — but the building itself has not changed.
Case study: a house rated G in a slack market
To make the discount concrete, let's take an illustrative case — the amounts are indicative and do not replace a local valuation or a simulation. You are considering selling a 110 m² house in a rural area, rated G, whose "all-classes" market price would stand at around €200,000 for an equivalent D-rated property in the area.
Scenario — Discount on a 110 m² G house (average Notaires gap applied to a reference D price)
| Item | Detail | Amount |
|---|---|---|
| Reference price (equivalent class D) | Local market price | €200,000 |
| Average house G vs D discount | -25% (Notaires, 2024 transactions) | -€50,000 |
| Expected price as-is (class G) | €200,000 − €50,000 | €150,000 |
| Gross shortfall linked to the label | vs reference D price | €50,000 |
The result illustrates the stakes: a theoretical €50,000 gap between a G and a D on this profile. The rational decision is not mechanically "renovate": you must weigh this shortfall against the cost of the works needed to climb a class, the available subsidies and the timeline. In a slack market where the price per square metre is low, a heavy renovation may cost more than the discount it erases; in a tight market, the trade-off often reverses. This is exactly what a personalised calculation lets you settle.
The interpretation mistakes that cost dearly
Mistake no. 1 — Applying the national discount to your property without accounting for the region
The "-25% for a G house" discount is a national average. In a tight market, the real gap can be two to three times smaller; in a slack rural area, it can exceed it. Blindly applying the average leads to undervaluing a property in a tight market or overvaluing it in a slack one. Always cross-reference class, property type and location — which is exactly what the Notaires' model retains.
Mistake no. 2 — Confusing listing price and sale price
The strength of the Notaires' study is that it rests on completed sale prices, not on listings. Yet sieves are often heavily negotiated down between listing and deed: a seller anchoring on the displayed prices of high-performance homes overvalues their property and freezes their sale. The discount materialises at signing, not in the listing.
Mistake no. 3 — Ignoring the 2026 DPE reform
Estimating a "sieve" discount in 2026 without checking the class applicable after the electricity coefficient reform risks a misreading. An electrically heated home may have changed class on 1 January 2026 with no works. Conversely, an oil- or gas-heated home did not benefit from this effect. Check the current label before any calculation.
⚠️ Warning: never reason solely from the "the discount is worsening" trend to postpone a sale. While the approaching 2028 rental ban (class F) pushes sieves down, waiting can also mean selling into a market where sieve supply is piling up. The exit window for an inefficient property narrows as the deadline approaches.
Mistake no. 4 — Forgetting the cost of use for the buyer
The discount does not only reflect the value of the works: it also prices in the future cost of use (high energy bills) and regulatory risk. Presenting a buyer with a clear diagnosis, a costed works quote and an achievable class trajectory reduces their uncertainty — and therefore the discount they demand. A "sieve but documented" property defends itself better than an opaque one.
Estimate the real discount on your property
Estimate the impact of the energy label on your home's value: the simulator cross-references the DPE class, the property type (house or flat), the local market price and your municipality to compute a discount range — aligned with the logic of the Notaires' green value — and weighs it against the cost of a renovation and the class gain targeted.
To go further: the DPE discount, city by city and the 2028 rental shock in sieve cities.
Conclusion
The Notaires de France "green value" study establishes a now-undeniable fact: the discount on thermal sieves is real, measured on completed sales, and it has worsened every year since 2021 — reaching -25% for a G house and -12% for a G flat on 2024 transactions. The regulatory trajectory (rental bans staggered through to 2034, the EPBD directive) makes a continuation of this worsening highly likely. As soon as the 2025 edition is published, these orders of magnitude will need updating — factoring in the side effect of the 2026 electricity coefficient reform, which has already re-rated part of the stock.
To move from national statistics to your own situation, the OneDpe DPE discount simulator quantifies the gap applicable to your property by class, type and location, and weighs it against the cost of a renovation — to decide, for example ahead of the class-E rental ban, drawing also on the rules and exceptions of the 2034 ban.
