Statistical Models for Valuation

Definition

Statistical Models for Valuation are financial models that use granular data about assets and transaction prices to establish a Valuation appraisal of a particular asset / property. They are one of the potential methodologies.

EBA Requirements

Institutions should set out, in their policies and procedures, the criteria for using statistical models for the purposes of valuation, revaluation and monitoring the values of collateral. These policies and procedures should account for such models’ proven track record, property-specific variables considered, the use of minimum available and accurate information, and models’ uncertainty.

Institutions should ensure that the advanced statistical models used are:

• property and location specific at a sufficient level of granularity (e.g. postcode for immovable property collateral);
• valid and accurate, and subject to robust and regular backtesting against the actual observed transaction prices;
• based on a sufficiently large and representative sample, based on observed transaction prices;
• based on up-to-date data of high quality.

When using statistical models, institutions are ultimately responsible for the appropriateness and performance of the models, and the valuer remains responsible for the valuation that is made using a statistical model. Institutions should understand their methodology, input data and assumptions of the models used. Institutions should ensure that the documentation of models is up to date.

Institutions should have adequate IT processes, systems and capabilities in place and sufficient and accurate data for the purposes of any statistical model-based valuation or revaluation of collateral.

See Also

• EBA Guidelines on Loan Origination and Monitoring