This paper presents a framework for conducting effective stress tests and incorporating insights from stress tests in portfolio construction. We examine how to determine the scope of the test, how to construct severe, but plausible scenarios, how to transmit the shock to the portfolio and how to incorporate the results of stress tests in portfolio construction. Stress testing can be a useful complement to risk model outputs, such as volatility, VaR, and expected shortfall. The key advantage of stress tests is that the loss is linked to a specific event, which can be more meaningful to portfolio managers than a summary statistic of a loss distribution. Prior research on stress testing has concentrated on ways to develop realistic and relevant shocks. The framework presented here attempts to expand on this, by illustrating that stress testing is a broader process addressing a wide range of investment problems and is useful in all stages of investment decisions.
Publication: Research Insight
Authors: MELAS Dimitris, RUBAN Oleg
