STRESS TESTING & MODEL RISK
Ask what the base case leaves out.
Stress testing and model risk research examine adverse scenarios, assumption sensitivity and the limits of historical relationships.
01 / Eutaxion
Move beyond a single expected outcome.
A base case is a useful organizing device, but it represents only one set of assumptions. Stress testing examines how a decision may behave when those assumptions are challenged.
Scenarios can help connect changes in economic conditions with their possible consequences for exposures and liquidity.
02 / Eutaxion
Design scenarios with a purpose.
A scenario should address a meaningful vulnerability or research question. It may examine a historical type of disruption, a hypothetical combination of pressures or a specific assumption on which a result depends.
The value lies in the reasoning it exposes, not in presenting the scenario as a prediction.
03 / Eutaxion
Recognize model dependence.
Model risk includes the possibility that an analytical framework is misspecified, uses unsuitable data or is applied outside the conditions for which it was designed.
Sensitivity analysis and comparisons with alternative methods can help reveal how strongly a conclusion depends on particular choices.
04 / Eutaxion
Translate findings into review.
Stress results should inform the questions asked of a decision and the conditions that warrant reassessment. A model’s limitations need to be communicated alongside its outputs.
No set of scenarios can cover every possible outcome. The process should remain open to risks that were not anticipated.
Continue exploring
A connected perspective.
A considered conversation