The Art Of Designing A Stress Test For Market Risk: A Step-By-Step Approach
The days when confidence levels in single risk metrics like VaR and CVaR were high came to an abrupt end on September 15, 2008, when Lehman failed. Almost a decade later, the practice of combining statistical models with stress tests has become the norm, but designing the right stress test remains as much art as science. Risk managers looking to leverage the power of stress tests are facing a range of choices. Which stress test - historical, user-defined or transitive? Which risk factors? This blog outlines a step-by-step approach to creatively designing meaningful stress tests for market risk.
Step 1: Select the Right Kind of Stress Test
Selecting the right stress test for your portfolio is as fundamental to risk management as selecting the right canvas is to painting a masterpiece– get it wrong, and all subsequent considerations are for naught. The four basic types of stress test are:
- Historical stress tests
- User-Defined Stress Tests
- Transitive Stress Tests
- Partitioned Transitive Stress Tests
Each one has pros and cons that need to be considered in the context of the portfolio in question:
Step 2: Fine-tune Your Inputs to Make it Meaningful
Designing a stress test requires sketching out possible scenarios. This in turn helps to select the most appropriate risk factors in relation to the portfolio, be they pricing factors or risk factors:
Choose the right factors
- Pricing factors: Risk or portfolio managers who are interested in short-term risk measures on an actively traded portfolio will choose risk factors that are very closely aligned with the pricing factors, such as bond prices, price returns for each stock in the portfolio, and so on.
- Risk factors: Those who need to design stress tests for portfolios with longer investment horizons will choose less granular factors, such as fundamental factors aligned with their overall investment outlook, such as key rate duration at the 6-month, 1-5-10- and 20 year nodes.
Step 3: Apply Best Practice to Your Stress Tests
While designing a meaningful stress test is a creative task with many dimensions, the following guidelines should be included:
- Use historical stress tests, but only if you believe history will repeat.
- Use transitive and user-defined stress tests together.
- Update stress tests on a regular basis to reflect changing market conditions and portfolio composition.
- Use multi-horizon with instantaneous stress tests, as one can observe significant differences between instantaneous and multi-horizon stress tests during periods of pronounced implied volatility.
- Employ diagnostics to assess the plausibility of the size of shocks.
Stress tests are both art and science, but require a systematic approach that takes into account the specifics of your portfolio. A large degree of creative thought and flexibility is required to ensure the right risk factors are applied to the right risk model, but there needs to be transparency and consistency in the approach, which can be supported with the right tools.
To learn more, read Axioma Insights: Stress Testing Best Practices here.