Risk, policy, and models - committee language
Policy (rules)
- What it is: “If this, then that” decisions your bank wrote down (caps, prohibited segments, required documents).
- Why committees like it: Deterministic, easy to cite in a memo.
- Where it shows: Rules engine and decision reasons on loan screens.
Models (statistics)
- What it is: Patterns learned from history to estimate risk or behavior-outputs are probabilities or scores.
- Why model risk cares: Needs validation, drift checks, version control.
- Where it shows: Scoring, default prediction, parts of dynamic pricing, ML center diagnostics.
Dynamic pricing sits in the middle
Pricing combines model outputs, rules, and regulatory ceilings (for example NBE ranges) to propose rates or bands-still subject to human approval where your policy says so.