Don't Stress Out Over Stress Tests

In the wake of Lehman Brothers’ failure in 2008, the meaning of “stress test” changed. The worst global financial and economic crisis in the last 80 years created a confidence gap that forced regulators and institutions into simulating worst-case scenarios.

In two big waves, 2009 for USA and 2010 for Europe, the majority of the global financial system was tested to the top. This stressed the risk managers more than the portfolios.

Now that stress test fever is back, what have we learned that can help us get through? In a series of posts, I will discuss how to be prepared for stress tests, what the market learned and how to apply those lessons to the day-by-day rigors of risk management.

The first lesson is: Don’t panic. Stress tests are inevitable, so relax and enjoy! We know regulators will demand these on a regular basis, so create a proper framework for testing your portfolios.

Build a simulation environment, your very own Stress Testing Lab, with the proper software tools and data sets. There are several types of tests that can be applied to a specific portfolio, the most important for risk managers being the credit risk stress tests that reflect potential losses from defaults on loans, including consumer loans and corporate loans. Stress tests for credit risk examine the impact of rising loan defaults, or non-performing loans, on bank profit and capital.

In simple terms: It's a what-if analysis on how changes in exchange rates, interest rates, housing prices, unemployment, GDP growth and other variables affects the value of the assets in a bank’s portfolio, as well as its profits and capital. Complexity can vary, ranging from simple sensitivity tests to complex stress tests that assess the impact of a severe macroeconomic stress event on measures like earnings and economic capital.

Your Stress Test Lab should include a statistically valid sample of the portfolio (if not all your data), and decision engines, such as FICO DMP, that will simulate the effects of the proposed scenarios in your strategy, as well as analytical models and policies to define the future capital needed  and the expected return on that capital. Such a tool must have a framework that includes statistical capabilities to evaluate impacts and perform simulations, a decision engine to run the test itself, and a comprehensive reporting tool to help describe the results for expert evaluation and create the final executive dashboard. And, of course, the framework must include a case manager to facilitate expert intervention in the whole stress test process.

Fortunately, user-friendly new technologies are available for your Stress Test Lab, such as the FICO Decision Management Platform. It is an elegant way to prepare you for the regulators’ demands. And also to reduce your own stress!

In the next post, let’s discuss what methodologies can be applied to the Stress Lab Testing.