Guides
How property valuation, hazard risk and price data actually work
Definitions, methods and source inventories written for the people who buy and audit this data: insurers, lenders, investors and the analysts who check their numbers. Every page opens with the answer, carries its figures in tables, and lists the public sources it rests on.
Valuation
Banks and regulation
- What CRR3 changed in how banks value mortgage collateralWhat Article 229(1) requires of a property value, and how the revaluation cap works on one flat.
- Can a bank use an AVM under CRR3?What Article 208(3a) and the EBA guidelines let a model do with mortgage collateral, and what to ask an AVM vendor.
- How AVM accuracy is measuredThe accuracy measures of the European AVM Alliance standards, and the results of our own test against asking prices.
- Which sale prices are public in Slovakia, Czechia, Poland and HungaryWhat sale-price data each market publishes, and what a lender can back-test a model against.
- Is an AVM high-risk under the EU AI Act?How Annex III point 5(b) reads for a model that values property, and when the high-risk obligations apply.
- How often banks must revalue collateral under CRR3How often collateral must be monitored, and why the monitoring cadence now limits an upward revaluation.
- National rules on AVMs in mortgage lending in Slovakia, Czechia, Poland and HungaryWho may value mortgage collateral in each market, and where a statistical valuation can carry a new loan.
- Glossary of bank collateral valuation termsCollateral valuation terms, each with the legal or standards text that defines it.
Insurance
- Rebuild cost vs market value: how the sum insured is setWhy the two values diverge, how carriers set the sum insured, underinsurance and proportional reduction, with a worked example.
- LiDAR building volume as a rebuild-cost inputHow measured building volume replaces footprint-times-storeys guesses in reconstruction cost, and where it breaks.
Risk
- Expected annual loss for a property, with a worked exampleExceedance probability, return periods, depth-damage curves and how the annual loss turns into a technical premium.
- How flood models workThe seven steps from rain to an insured loss, river floods and rain floods step by step, return periods, river flow and what a test against a real flood proves.
- Address-level flood risk scoring: zones, return periods and depthFluvial vs pluvial flooding, EU Floods Directive maps, modelled depth vs zone membership and why district scores fail.
Methodology
Data sources
- Flood-risk data sources in SlovakiaSVP hazard and risk maps, SHMÚ hydrology, ÚGKK elevation data and the European JRC maps: what each holds and how to get it.
- Flood-risk data sources in CzechiaČAP insurer flood zones, DIBAVOD and ČHMÚ data, state river basin enterprises and the European JRC maps.
- Flood-risk data sources in PolandISOK hazard maps, Wody Polskie, IMGW, GUGiK elevation data and the European JRC maps.
- Flood-risk data sources in HungaryOVF hazard and risk maps, the Danube and Tisza defence system, Lechner elevation data and the European JRC maps.
- Property price data sources in SlovakiaNBS price statistics, the statistical office, Eurostat, the cadastre and portal asking prices: what is public and what is not.
- Property price data sources in CzechiaČSÚ transaction indices, ČNB, Eurostat, the cadastre price records and portal asking prices.
- Property price data sources in PolandThe NBP housing market report, GUS, Eurostat, county price registers and portal asking prices.
- Property price data sources in HungaryMNB and KSH indices, Eurostat, the paid cadastral registers and portal asking prices.
How these pages are written
- Answer first. The opening paragraph is the complete short answer; the rest of the page is the working.
- Every number traces to a public source or a named dataset, listed at the end of the page. A figure that cannot be sourced is described, not invented.
- Bytero appears as one worked example of a method, not as the subject. Where availability matters, the page points to the coverage page rather than making a claim in passing.
- Pages are revised when the data behind them changes, not re-dated for the sake of it. The vintage of each figure is stated in the sentence that uses it.
Also on this site: vendor comparisons, the glossary, the Bytero Live asking-price index and the flood check by address.