Skip to content

Declared loss model and estimated loss distribution

A declared scenario model supplies probabilities; an estimate needs observations and an estimation method.

On this page

Read the original teaching model

The public tool’s default −10/0/10/50 outcomes and 30/40/25/5 percent probabilities are stipulated. They let readers verify quantiles and tail allocation. They are not empirical probabilities for a named portfolio.

Retain evidence for an estimate

Input-status comparison
StatusEvidence needed
Declared teaching scenarioExplicit losses and probabilities
Historical empirical estimateDates, observations, weighting and quantile rule
Simulation estimateDistribution, exposures, model and generated outcomes
Stress scenarioSpecified adverse factor changes and repricing basis

Identify what arithmetic checks establish

Probabilities summing to one and an internally consistent allocation prove that the selected finite-model calculation can be performed. They do not validate missing observations, regime stability, pricing assumptions or future calibration.

Keep unresolved data questions visible

Record source, horizon, return/loss conversion, estimator and unresolved limitations beside the result. If an estimate is unavailable, label the numbers as assumptions rather than supplying invented historical dates.

Further references