Stress Testing

In risk management, the notion of "stress test" refers to some extraordinary situation occurring very rarely but whose consequences would be dramatic for a given portfolio. Such situations are usually outside the scope of normal market conditions, but they need to be envisaged and their consequences need to be understood.

Stress testing therefore helps hedge fund managers to determine how their portfolio would react in stylized scenarios. It gives them a better understanding of where extreme risks lie in their portfolios, and allows them to prepare so that they are able to act more decisively and more quickly if the worst-case takes place unexpectedly.

Measuring the volatility and/or value at risk (VaR) of a portfolio provides objective measures, which are usually based on some statistical observation of the past. By contrast, stress testing is a subjective risk-measurement approach that depends mainly on human judgment and experience.

At its simplest, a stress test will show the sensitivity of a portfolio to a certain change in some underlying risk factors. These changes (called "scenarios") can be based on historical data (October 1987, summer 1998, etc.) or can be hypothetical and entail large movements considered being possible.

Stress tests are helpful for evaluating the ef ects of large movements in key variables. Hedge funds ot en use them as a complement to statistical models such as VaR to capture the impact on a portfolio of exceptional but plausible large loss events, understand the overall risk proi le of a fund, set limits, and take capital allocation decisions. However, one should also be aware of the limits of stress test models.

In particular, they usually assume that the portfolio stays unchanged over the stress test period, and are often not able to capture the entire spectrum and interplay of risk exposures (such as operational risk, legal risk, liquidity risk, etc.). As an illustration, many hedge funds run a summer 1998 scenario on their portfolio but they do not model the lack of liquidity associated with this crash.

Stress Testing
Stress Testing