Stress Testing for Your Portfolio

Stress Testing for Your Portfolio

Stress Testing for Your Portfolio

Stress testing is a risk management method to evaluate the impact on portfolio values of potentially negative events. Gloabl Association of Risk Professionals (GARP) recently released a post “COVID-19 Calls for New Risk Measures” highlighted some problems of our traditional risk assessment. The COVID-19 financial turmoil should lead financial professionals to look into their investment portfolios and have a more accurate stress testing since some investment and model portfolios are not prepared to deal with similar risks.

Some known methodologies for stress testing for portfolio management include the use of Value at Risk (VaR) and Extremem Value THeory (EVT). Value at risk focuses on the distribution of risks over the entire distribution curve while the Extreme Value Theory focuses on the shaping the tails. For most stock portfolios, we use VaR because portfolio managers typically are more concern how to flatten the curve rather than hedging certain risks in their investment portfolios.

Stress testings can involve the use of univariate and multivariate stress tests. Univariate test are simple and easy to conduct but its limitations are also apparent because the test might ignore input dependencies and thus leading to collinearity, causing some independent regression coefficients to be wrong, or curve fitting. Multivariate stress tests solve such problems and are recommended for multi-asset portfolios.

Both methods should be used with scenario analysis for more realistic results. Most backtesting tools on the market are based on historical events; but in reality, risk management should be also applied with hypothetical scenarios as well. The financial crisis of 2008 resulted in the lack of understanding of risk dependencies between the risk levels of multiple variables. Another scenario is the use of hybrid scenarios, which use both historical and hypothetical scenarios for  your risk analysis.

We model the distribution of returns with GARCH model combined with VaR for portfolio stress testing. Depending on your portfolio management style and active levels, we might recommend different methods. If you have questions regarding the use of technology and financial model for your portfolios, Contact Us and learn about stress testing for your portfolio.