Summary Stress testing produces value when it changes decisions. Programmes that produce polished packs but no observable change in capital plans, appetite or business strategy have failed at their purpose.
Scenario design Scenarios should be severe but plausible, and tailored to the firm's specific vulnerabilities. Standardised regulatory scenarios are a floor, not a ceiling.
Reverse stress testing Reverse stress asks what would break the firm and works backward. It is uniquely useful for surfacing tail vulnerabilities that forward scenarios miss.
Integration Stress results should feed ICAAP, ILAAP, recovery planning, appetite reviews and business planning. Stress that lives only inside the risk function is not integrated.
Management actions Assumed management actions should be realistic, timed and evidenced. Actions that require deep and functioning markets in stress are not credible.
Governance Scenario approval, model use, results challenge and communication to the board should follow a documented cycle with clear ownership.
Limitations Stress testing is a structured argument, not a forecast. Its purpose is to illuminate exposures and inform decisions under uncertainty.
Related expertise See [Enterprise Risk Management](/expertise/enterprise-risk) and [Banking Risk](/expertise/banking-risk).
Frequently asked questions
What should risk leaders know about scenario design?
Scenarios should be severe but plausible, and tailored to the firm's specific vulnerabilities. Standardised regulatory scenarios are a floor, not a ceiling.
What should risk leaders know about reverse stress testing?
Reverse stress asks what would break the firm and works backward. It is uniquely useful for surfacing tail vulnerabilities that forward scenarios miss.
What should risk leaders know about integration?
Stress results should feed ICAAP, ILAAP, recovery planning, appetite reviews and business planning. Stress that lives only inside the risk function is not integrated.
What should risk leaders know about management actions?
Assumed management actions should be realistic, timed and evidenced. Actions that require deep and functioning markets in stress are not credible.
What should risk leaders know about governance?
Scenario approval, model use, results challenge and communication to the board should follow a documented cycle with clear ownership.