Skip to content

Historical loss record review

Keep data construction separate from the arithmetic applied to an empirical loss table.

Omni Finance AcademyBy Omni Finance Academy
On this page

Retain the estimation evidence

The worksheet has review fields for loss dates, current exposure basis, factor changes, repricing or approximation, missing observations, weighting, horizon and percentile rule. Blank entries indicate unresolved evidence, not an approved estimate.

Identify what historical losses describe

Past realised returns on a changing portfolio may differ from losses obtained by applying historical factor moves to today’s positions. Record which construction is being used and any nonlinear or omitted exposure treatment.

Name the sample method

Empirical review fields
FieldReview question
Observation weightingEqual or explicitly unequal probabilities?
QuantileLower empirical crossing or interpolation?
Tail allocationWhole or fractional boundary mass?
HorizonMatching period for every record?
Missing dataExclusion, imputation or another method?
Regime/sourceWhat observations support the estimate?

Keep the download’s limits clear

This CSV is a review record, not a repricing engine, statistical estimator or evidence that future losses are covered. Obtain missing source information before turning a scenario result into an empirical claim.

Further references