Glossary

Plain-language definitions of the terms this bundle uses. It grows as the bundle does.

Model risk. The risk that a model is simply wrong — not misused, not misaligned with policy, just producing inaccurate estimates — and that being wrong causes a financial or operational loss.

Model. A complex quantitative method, system, or approach that applies statistical, economic, or financial theories to process input data into quantitative estimates. Excludes simple arithmetic, spreadsheets, and deterministic rule-based processes.

Conceptual soundness. Reviewing a model's design, construction, and developmental testing before it goes live — the pre-launch check.

Outcomes analysis. Comparing what a model predicted against what actually happened, via back-testing or outlier analysis — the reality check.

Ongoing monitoring. Continuously evaluating whether a live model still performs as expected as products, data, and market conditions change. Where drift gets caught.

Effective challenge. Critical review by experts who are both independent enough to be objective and senior enough to actually force a change. The governance concept that makes validation mean something.

Materiality tiering. Scoping how much scrutiny a model gets based on its exposure (how much is riding on it) and its purpose — not every model needs the same level of review.