IFRS 9 provisions: preparing data to calculate expected credit losses

Distinguish the calculation's purpose from the rules for applying it

The IFRS Foundation describes IFRS 9's impairment model as an assessment of expected credit losses. It requires a forward-looking approach rather than waiting for a loss to occur. First, however, identify which financial instruments and requirements apply to the particular organisation's reporting.

Do not conflate IFRS accounting estimates, prudential calculations and tax consequences. Their purposes, bases and rules may differ. A practical recommendation is to record the purpose of each calculation and separately document the reconciliation between its result and the relevant reporting.

Prepare a reproducible loan portfolio snapshot

Fix the calculation date and retain the original extract. For each field, describe its meaning, source and treatment of missing values. If an agreement has different numbers across systems, prepare a mapping table. Otherwise, payment history may be incorrectly linked to another obligation.

The data required depend on the model, but checks can start with the list below. It is a practical preparation list, not a universal form prescribed by the standard. Unexplained gaps and duplicates should be disclosed in a data quality report before calculation.

Support the risk assessment and loss horizon

Under the general IFRS 9 model, a significant increase in credit risk is associated with moving to lifetime loss measurement. Twelve-month expected credit losses are not simply missed payments over the next twelve months: they are the portion of lifetime losses associated with defaults that may occur within that horizon.

For a working calculation, describe the criteria for changes in risk, contract grouping and assumptions used. Do not mechanically copy another organisation's parameters: products, customers, security and recovery history may differ. Where reliable history is limited, explicitly state the limitation and rationale for the approach rather than creating the appearance of precise statistics.

Check forward-looking assumptions and result sensitivity

The IFRS Foundation's publicly available guidance emphasises reasonable historical, current and forward-looking information and rejects a purely mechanical approach. Its 2020 material explains these principles in the context of uncertainty; it does not add new requirements to the standard.

It is useful to retain the source of each key assumption, the date selected and an explanation of its connection to the portfolio. Compare results under changes to major assumptions. This helps reveal whether the result is driven by a single unsupported coefficient. The number of scenarios and their weighting method should be justified by the organisation's methodology, not chosen to create an attractive table.

Prepare the calculation so another specialist can reproduce it

Retain source data, data-cleaning rules, the model version, parameters and segment-level results. Reconcile the total to accounting registers; explain changes from the previous date separately for portfolio movements and changes in estimates. When an error is found, correct it at the relevant stage, not only in the final cell.

A useful discussion package includes the methodology, data quality report, calculation, explanation of material changes and limitations. ProFinance Consulting provides provision calculation and accounting organisation services. The applicability of a particular methodology and scope of work are determined after analysing the portfolio and calculation objective.

Preparation checklist

• The calculation's purpose, date and scope of instruments are defined.
• The source portfolio is reconciled with accounting data.
• Missing values, duplicates and contract changes have been checked.
• Risk and segmentation criteria are documented and supported.
• Material forward-looking assumptions have identified sources.
• The result is reproducible, with limitations and changes disclosed.

Frequently asked questions