Stress Testing and Scenario Analysis for Portfolios
Standard risk metrics like VaR assume normal market conditions. Stress testing and scenario analysis reveal how portfolios behave in the extreme conditions that matter most — market crashes, liquidity crises, and regime changes.
Why Standard Risk Metrics Fail in Crises
Value at Risk (VaR) and other statistical risk measures are calibrated on historical return distributions. During normal markets, these distributions are approximately normal. During crises, they are not — correlations spike, liquidity evaporates, and losses far exceed what normal distributions predict. The 2008 crisis produced "25-sigma" events that VaR models said should occur once in the lifetime of the universe.
Stress testing complements statistical risk measures by asking: "What happens to this portfolio if X occurs?" — where X is a specific historical crisis or a plausible hypothetical scenario. The results are not probabilistic; they are conditional: given this scenario, here is the expected portfolio impact.
Historical Stress Scenarios
| Scenario | Equities |
|---|---|
| 2008 Global Financial Crisis | -50% |
| 2020 COVID Crash | -34% (peak-trough) |
| 1994 Bond Market Massacre | -5% |
| 2022 Rate Shock | -20% |
Hypothetical Scenario Design
Beyond historical scenarios, risk managers design hypothetical scenarios based on plausible future risks. Common hypothetical scenarios include: a rapid rise in inflation forcing central banks to hike aggressively; a geopolitical shock disrupting energy markets; a sovereign debt crisis in a major economy; or a sudden reversal of the AI-driven equity rally.
Good scenario design starts with a narrative — a coherent story about what happens and why — and then translates that narrative into specific market moves across asset classes. The narrative discipline prevents cherry-picking scenarios that only affect certain parts of the portfolio.
Translating Results into Action
Stress test results are only valuable if they drive decisions. The output should answer three questions: (1) Is the portfolio's worst-case loss within the institution's risk tolerance? (2) Are there specific concentrations or correlations that amplify losses in stress scenarios? (3) What hedges or portfolio changes would most efficiently reduce stress losses?
Many institutions establish formal stress loss limits — if a specific scenario would produce a loss exceeding X% of assets, the portfolio must be adjusted. This creates a direct link between stress testing and portfolio construction, ensuring that risk management is not just a reporting exercise.