Stress Testing

Study guides: Exam 6C, Exam CFE-101, Exam CP-351

Stress Testing is the examination of an insurer's financial position under severe but plausible adverse conditions. It is the general technique underlying FCT and ORSA, and OSFI expects it to be embedded in risk management rather than performed once a year for a report.

• Sensitivity test — move one variable and observe the effect (a 100 basis point rate change, a 10\% reserve deterioration). Useful for identifying which assumptions matter, and for calibrating attention.
• Scenario test — move a coherent set of variables together in a way that could actually happen: a recession raising claim frequency and fraud while depressing asset values and premium volume. Scenarios are harder to build and far more informative, because real adversity is correlated.
• Reverse stress test — start from failure and work backwards: what would have to happen for this insurer to become non-viable? It surfaces vulnerabilities that forward scenarios, anchored on the risks management already worries about, systematically miss.
• Severity calibration is the hard judgement. A test too mild reveals nothing; one too extreme is dismissed as implausible and ignored. The standard is severe but plausible, and the calibration should be documented — usually by reference to historical events, industry experience, or a stated return period.
• Ripple effects and management actions must both be modelled: the scenario's second-order consequences, and the realistic responses available. Assuming management can raise capital in the middle of the scenario that caused the problem is the classic way to make a stress test useless.
• The output must reach a decision. A stress test that produces a report and no change to limits, reinsurance, capital targets or the business plan has not been used. OSFI's supervisory reviews look for the link between test results and action.

Sensitivity, Scenario and Reverse An insurer wants to understand its exposure to an economic downturn. Design one test of each type and explain what each reveals.

Answer Sensitivity test. Reduce equity values by 25\%, holding everything else constant. Result: capital available falls by the equity holding times 25\% times (1-t); the MCT ratio falls accordingly.

What it reveals: how much of the capital ratio depends on equity values. Clean, quick, and it tells management whether equities are worth further attention. What it misses: everything else that happens in a downturn.

Scenario test. A recession over two years: equities -25\%; credit spreads widen, so corporate bond values fall and one holding defaults; claim frequency rises 6\% in commercial lines and fraud indicators rise in personal auto; premium volume falls 8\% as insureds reduce coverage; and reserve development turns adverse as claimants pursue larger settlements.

What it reveals: the correlation. Each element alone is survivable; together they attack capital available (investment losses, underwriting losses) and the Base Solvency Buffer (adverse development raising liabilities) simultaneously. This is where the Diversification Credit assumption gets tested, and it is the test that changes decisions.

Reverse stress test. Ask: what combination would take the MCT ratio below 100\%? Suppose the answer is a 1-in-250 catastrophe combined with the failure of the insurer's largest reinsurer.

What it reveals: the actual failure mode — which turns out to be reinsurer concentration, not investment risk at all. Forward scenarios built around market and underwriting risk would never have found it, because nobody was worried about the reinsurer.

The lesson. The three are complements, not substitutes. Sensitivities size individual exposures, scenarios test correlation, and the reverse test finds the vulnerability management has not thought of — which is usually the one that matters.