DVF Data in Paris: How to Read the Real Sale Prices

DVF data (Demandes de Valeurs Foncieres) is the most powerful and least known tool for property buyers in Paris. Published by the Direction Generale des Finances Publiques, it reveals the real sale prices: not the prices listed by agents, not the estimates from property portals, but the amounts actually paid at the notary’s office. And the difference between the two is considerable: in Paris, the average gap between listed price and actual sale price is 8% in 2026, roughly 40,000 euros on a property worth 500,000 euros.

What exactly is the DVF database?

The DVF database records all property transactions completed in France, transmitted by notaries to the land registry office. Since 2019, this data has been open data, freely accessible to anyone. Before that date, only property professionals and notaries had access. It was a quiet revolution in the transparency of the French property market.

Each transaction recorded in DVF contains a precise set of information: the date of transfer, the exact address of the property (number, street, postcode, municipality), the nature of the transfer (sale, auction, expropriation), the total sale price, the actual surface area (surface Carrez for co-ownership lots), the property type (apartment, house, land, commercial premises), and the number of main rooms.

For Paris, the database contains approximately 35,000 to 40,000 transactions per year, a statistically massive sample that enables detailed analysis by arrondissement, by neighbourhood, and even by street. It is this depth that makes DVF an incomparable tool.

How to access DVF data

Three official access points exist, each with its own advantages.

The DVF map application (app.dvf.etalab.gouv.fr) is the most accessible entry point. An interactive map lets you click on any plot of land in France and display all transactions from the past five years. For each sale: the date, the price, the surface area and the property type. It is the ideal tool for a buyer who wants to know the real prices on a specific street. No registration, no account creation: access is immediate and free.

The data.gouv.fr portal provides access to downloadable raw files. This is the option for those who want to analyse the data in depth, for example to calculate the median price per sqm in a neighbourhood over three years, or to identify price trends street by street. The files are in CSV format, usable in Excel or any data analysis tool. This is the format we use at Home Select for our market analyses.

The Notaires de France website (immobilier.statistiques.notaires.fr) offers a more polished interface with comparison tools and pre-formatted market indicators. The data is the same as DVF (it comes from the same notarial deeds) but the presentation is more digestible for an individual buyer. The tool allows you to compare the price per sqm of a neighbourhood with the average for the arrondissement and for Paris.

Key figure: In 2026, the DVF database contains more than 280,000 Parisian transactions over the five most recent consultable years (2020-2025). This is a volume of data without equivalent in any other European country: France is one of the few states to publish all its property transactions as open data.

The revelation: the gap between listed price and actual price

This is where DVF becomes a formidable negotiation tool. By cross-referencing listed prices on property portals with actual sale prices recorded in DVF, you can measure the systematic gap between what the seller asks and what the buyer ends up paying.

In Paris, this average gap is 8% in 2026. But it varies considerably depending on the arrondissement and the property type.

In the most sought-after arrondissements (6th, 7th, 4th), the gap is moderate: 3 to 5%. Demand is strong, properties sell quickly, and sellers adjust their asking prices more accurately. A property listed at 15,000 euros/sqm in the 7th typically sells for between 14,300 and 14,550 euros/sqm.

In the mid-range arrondissements (9th, 10th, 11th, 14th, 15th), the gap widens: 6 to 9%. Sellers often set a “test” price to gauge the market, and negotiation is an expected part of the process. A three-bedroom apartment listed at 650,000 euros in the 11th sells on average around 600,000 to 610,000 euros.

In the least expensive arrondissements (13th, 19th, 20th), the gap can reach 8 to 12%. Buyers there are more sensitive to credit conditions, selling times are longer, and sellers often end up accepting significant reductions. A property listed at 8,500 euros/sqm in the 19th can sell at 7,600-7,800 euros/sqm after several months on the market.

By property type, studios and one-bedroom apartments show the largest gaps (9-11%): the small unit market is the most volatile and most sensitive to credit conditions. Family three-bedroom apartments in good locations have the smallest gaps (4-6%): demand structurally exceeds supply in this segment.

Jean Mascla’s advice: The most common mistake buyers make is to consider the listed price as the market price. DVF proves the opposite: the listed price is the starting point of a negotiation, not the finish line. A buyer who does not negotiate pays on average 8% more than the market, roughly 40,000 euros on a property worth 500,000 euros. That is the cost of a fitted kitchen, renovation works or two years of co-ownership charges.

How to read DVF data like a professional

Raw DVF data can be misleading if you do not know how to interpret it. Here are the five most common pitfalls and how to avoid them.

Pitfall 1: grouped lots. When a buyer purchases an apartment together with a cellar and a parking space, DVF sometimes records the transaction as a single lot at the total price. The price per sqm calculated on the apartment surface alone is then artificially inflated by the value of the cellar (3,000-8,000 euros) and the parking space (25,000-60,000 euros depending on the arrondissement). To isolate the true price per sqm of the apartment, you need to subtract the estimated value of the annexes.

Pitfall 2: atypical properties. A 150 sqm loft on the ground floor of a former workshop in the 11th cannot be compared to a 150 sqm Haussmann apartment on the 4th floor in the 7th. DVF does not mention the floor, the orientation, the condition of the property, or the building type. Two transactions at the same price per sqm on the same street can involve radically different properties in terms of quality.

Pitfall 3: the publication delay. The most recent DVF data is 6 to 9 months old. In a market moving at 2% per year, this delay is manageable. But in a rapidly shifting market, such as 2022-2023 when interest rates surged, DVF data may reflect a market that no longer exists. You should always contextualise the data relative to the transaction period.

Pitfall 4: the surface area. DVF indicates the actual surface area (surface Carrez in co-ownership), but errors exist. Some transactions show clearly incorrect surfaces: a two-room apartment of 12 sqm or a studio of 180 sqm. These are data entry or transmission errors. Any aberrant price per sqm (below 5,000 euros or above 30,000 euros, excluding ultra-premium addresses) should be verified.

Pitfall 5: sales between relatives. Family transfers, life annuity sales and court-ordered auctions are also recorded in DVF. Their prices do not reflect the open market: a family sale can go through at 30-50% below the market price. On a small sample (a quiet street with three transactions in two years), a single family sale can completely skew the average.

How Home Select uses DVF

At Home Select, DVF is one of the pillars of our valuation methodology, but never the only one. Our approach systematically cross-references three data sources.

DVF for the historical real price. For every property we evaluate, we start by analysing all DVF transactions within a 200-metre radius over the past 24 months. We filter out atypical sales (grouped lots, family sales, aberrant surfaces) and calculate a contextualised median price per sqm, by floor (DVF data does not indicate the floor, but we know the buildings), by estimated condition and by property type.

Our own transaction data. Since 2011, we have supported more than 1,200 transactions in Paris. For each one, we know the initial listed price, the final sale price, the negotiation margin achieved, the floor, the orientation, the condition of the property and the reasons for any discount or premium. This proprietary database, which we do not publish, is our competitive advantage. It allows us to contextualise DVF data with a level of granularity that public data alone cannot provide.

Real-time field observation. Our 16 property hunters collectively visit several hundred properties per month. They know the state of the market street by street, building by building. They know that a property on the market for 90 days on a particular street is negotiable by 8-10%, or that a renovated three-bedroom apartment in a particular building will sell at asking price in under two weeks. This field intelligence, impossible to capture in a database, is what transforms data into decisions.

Key figure: By cross-referencing DVF with our internal data, we estimate that our valuation of a property is reliable to plus or minus 3% of the final sale price. Online estimation portals, which rely solely on DVF and algorithms, show a margin of error of plus or minus 10 to 15%. This difference in precision, 3% versus 10-15%, represents 35,000 to 60,000 euros on a property worth 500,000 euros.

DVF as a negotiation weapon

The most concrete use of DVF for a buyer is negotiation. Here is how to turn data into arguments.

Step 1: build a file of comparables. Before making an offer, identify in DVF the 5 to 10 closest transactions: same street or neighbouring streets, comparable surface area, within the past 12 to 18 months. Calculate the median price per sqm. This is your market reference.

Step 2: measure the gap. Compare the price asked by the seller to the median price of your comparables. If the property is listed 10% above the DVF median price, you have a factual argument to negotiate. If the property is at or below the median price, the price is reasonable and negotiation will focus on other aspects (timeline, conditions, furniture).

Step 3: present the data. When formulating your offer, accompany the amount with a document presenting the comparable transactions extracted from DVF. It is factual, verifiable and indisputable: the seller and their agent cannot contest real sale prices recorded by the tax authorities. This document transforms the negotiation from an emotional power struggle into a rational discussion based on facts.

A concrete example: we recently assisted a buyer for a 68 sqm three-bedroom apartment in the 9th arrondissement, listed at 680,000 euros (10,000 euros/sqm). DVF analysis of transactions in the same area over 18 months showed a median price of 9,200 euros/sqm for comparable properties, giving a market value of around 625,000 euros. The property also had an EPC rating of E requiring insulation works estimated at 35,000 euros. Our offer of 610,000 euros, supported by DVF data and the works estimate, was accepted after a counter-offer at 625,000 euros. The buyer saved 55,000 euros compared to the listed price, exactly 8% negotiation, right on the Parisian average. Without DVF data, this negotiation would have been much harder to substantiate.

The limits of DVF, and why the human factor remains essential

DVF is a powerful tool, but it is only a tool. Several structural limitations prevent you from relying on data alone.

The 6 to 9-month delay creates a blind spot on the recent market. In a stable market, this is not a problem. But when conditions change rapidly, whether from a sudden rise or fall in interest rates, an economic shock, or a regulatory change, DVF data reflects a world that no longer exists. This is exactly what happened in 2022-2023: buyers relying on DVF prices from early 2022 were overvaluing the late 2023 market by 5-6%.

The absence of qualitative data is the other major limitation. DVF does not tell you whether the apartment was fully renovated or needed refreshing, on the 1st floor facing the courtyard or on the 5th with a view, in a renovated building or a struggling co-ownership. These factors can make the price of the same property vary by 20 to 30%. A DVF price per sqm without qualitative context is incomplete information.

Finally, DVF does not capture off-market transactions that do not complete, nor properties withdrawn from the market for lack of a buyer. The invisible market, the properties that could have sold but did not sell at the asking price, appears nowhere in the data. Yet this invisible market is an important indicator of the real level of demand in an area.

This is why human expertise remains irreplaceable. Data tells you what happened. A property hunter tells you what it means for your purchase, today, on this street, for this type of property. The combination of both, data and expertise, is what produces the best buying decisions.

Jean Mascla’s advice: Always consult DVF before visiting a property. Five minutes on the map application is enough to know the latest sale prices on the street. You will arrive at the viewing with a clear idea of market value, and you will avoid falling in love with a property overpriced by 15%. It is the most profitable habit a Parisian buyer can adopt.

---

DVF data is published by the Direction Generale des Finances Publiques (DGFiP) and accessible on data.gouv.fr. The analyses presented in this article are based on this data cross-referenced with our own observations from 1,200+ transactions supported since 2011.

Our property hunters cross-reference DVF data with 1,200 transactions to assess every property at fair value. Benefit from this data expertise for your property purchase in Paris. Speak to an expert