Today, investors can choose from a huge variety of different asset classes from various regions, making portfolio construction very complex. In doing so, they are not only concerned with maximizing return, but also with mitigating risk. The traditional mean-variance optimization framework by Markowitz uses the standard deviation of returns as the risk measure. However, the standard deviation does not reflect severe downside scenarios sufficiently. The conditional value-at-risk (CVaR) is a downside risk measure which focuses on the worst cases. In financial engineering, copulas are gaining interest. We bring these two concepts together to create a minimum CVaR portfolio in a multi asset setting. We apply a grouped t-copula to model the dependence structure between the returns. Using this model, we simulate the joint development of the yields and calculate the portfolio that minimizes the CVaR across the scenarios. By doing this, we include options already in the portfolio construction process to optimize the asset composition additionally with regards to hedging. To show the benefits of the presented scenario-based CVaR minimization framework we compare it to other common portfolio construction methods. It becomes apparent that the minimum CVaR portfolio exhibits superior downside risk characteristics while it still possesses a desirable upward potential.      «

Today, investors can choose from a huge variety of different asset classes from various regions, making portfolio construction very complex. In doing so, they are not only concerned with maximizing return, but also with mitigating risk. The traditional mean-variance optimization framework by Markowitz uses the standard deviation of returns as the risk measure. However, the standard deviation does not reflect severe downside scenarios sufficiently. The conditional value-at-risk (CVaR) is a dow...     »