Which sale prices are public in Slovakia, Czechia, Poland and Hungary

Why the gap matters

An asking price is what a seller hopes for. A sale price is what a buyer paid, usually after negotiation, sometimes after a long time on the market. A model trained and tested on asking prices learns the first. A bank's collateral rules are written for the second.

Three texts say so directly:

• CRR Article 208(3a)(b), added by CRR3: a monitoring model must be back-tested "against the actual observed transaction prices" and built on "a sufficiently large and representative sample, based on observed transaction prices".
• European AVM Alliance standards, section 4.6: "Asking Prices or estimated values that do not qualify as Surveyor Valuations are not considered as valid Benchmark Values."
• KNF Recommendation J (Poland, March 2023 version), point 2.3: the reliable sources for a bank's property database include notarial deeds, appraisal reports, the bank's own collateral assessments, mortgage lending values, land and building registers, land and mortgage registers, local plans, local transaction prices, statistics, construction cost catalogues and public authorities' price data. The list is open-ended, and it does not name asking prices.

What each market publishes

Poland

Poland changed the rules in 2025. An amendment to the Geodetic and Cartographic Law, published in the Journal of Laws on 12 November 2025 as item 1542, abolished the fees for data from the register of property prices. It took effect three months later, on 13 February 2026. Transaction prices can now be viewed and downloaded through the national Geoportal, without a fee or an application to each district office.

For aggregate figures, the central bank (NBP) publishes the BaRN database. It gives quarterly mean prices per square meter since the third quarter of 2006 for the 16 regional capitals and Gdynia, with offer and transaction prices for both the primary and the secondary market.

That makes Poland the one market of the four where the asking-to-sale gap can be read straight from an official source. For the second quarter of 2026, we divided BaRN's mean transaction price by its mean offer price per city:

On the secondary market the ratio was lowest in Gdynia, at 0.84, and highest in Lublin, at 0.99; Warsaw stood at 0.88. These are ratios of two means drawn from different flats, so they include differences in the mix of flats sold and offered. They are not a discount on a single flat. The gap still differs by more than 10 points between cities in the same quarter.

Polish banks also share collateral data through AMRON, an interbank database the banking sector has built since 2004. Recommendation J asks banks with a material mortgage book to take part in such a shared database.

Czechia

The Czech Statistical Office publishes two series built from cadastre data. CEN0402 gives the average purchase price per square meter for flats and for family houses, by region, district and size of municipality, for single years and three-year periods; 2025 is the latest year. CEN0401 is a quarterly index by region, to the fourth quarter of 2025.

A district average for the last full year is a useful check on a model's level. It cannot back-test single valuations, and comparing it with this year's asking prices mixes the price gap with a year of price growth.

Hungary

KSH table 18.2.2.14 gives the mean price per square meter by region and settlement type for used and new homes, quarterly. The table covers houses as well as flats, which matters when comparing it with a flat model.

Hungarian rules point to a second source. The decree on mortgage lending values (PM Decree 25/1997) refers to the sales value data supplied under PM Decree 33/2007 and builds its rules for statistical valuation on it. A lender may value a flat statistically for a new loan only where that data shows at least ten sales in the settlement in the previous calendar year, with unit prices no more than 30 percent apart. Our page on [national AVM rules](/guides/avm-mortgage-rules-sk-cz-pl-hu) sets out the full conditions.

Slovakia

The Slovak cadastre records the price of a property. Section 68(2) of the Cadastre Act (162/1995) says who may see it: the owner, and on request for their legal tasks, state bodies, Národná banka Slovenska, courts, the prosecutor, the police, notaries, bailiffs, the tax office, administrators under a special act, court-appointed experts in valuing property, farmland and forests, and "the person who prepares price maps". A bank is not on the list. A firm building a model would have to count as a person who prepares price maps, and the act does not say who does.

What is published is an index. The Statistical Office of the Slovak Republic releases a quarterly index of realized prices by region, computed from cadastre data. The central bank's statistics on residential property prices are built from asking prices. The agents' association Realitná únia keeps a register of closed transactions for its members and publishes aggregate figures.

So a Slovak model can be checked against the movement of sale prices by region, but not against their level, and not flat by flat. For a bank, the best set of sale prices in Slovakia is usually the bank's own, made of the purchase prices in its loan files and the full valuations by its panel.

What a lender can back-test against

In each of the four markets, the cleanest benchmark sits in a bank's own files, in the purchase prices of completed loans and the full valuations by its valuers. EBA paragraph 233 already asks banks to back-test their valuers; the same files can back-test a model.

How Bytero handles the gap

Bytero Engine is trained on asking prices, because they are published widely and without delay, while sale prices arrive late or not at all. It then converts an asking price into an estimated sale price using market signals such as time on market and the mortgage rate. Where official sale prices exist, the conversion can be checked against them; where they do not, it cannot be checked flat by flat.