Improved Conditional Value-at-Risk (CVaR) Based Method for Diversified Bond Portfolio Optimization

In this study, an improved CVaR-based Portfolio Optimization Method is presented. The method was used to test the performance of a diversified bond portfolio in providing low expected loss and optimal CVaR. A hypothetical diversified bond portfolio, which is a combination of Islamic bond or Sukuk and conventional bond, was constructed using bonds issued by four banking institutions. The performance of the improved method is determined by comparing the generated returns of the method against the existing CVaR-based Portfolio Optimization Method. The simulation of the optimization process of both methods was carried out by using the Geometric Brownian Motion-based Monte Carlo Simulation method. The results of the improved CVaR portfolio optimization method show that by restricting the upper and lower bounds with certain floor and ceiling bond weights using volatility weighting schemes, the expected loss can be reduced and an optimal CVaR can be achieved. Thus, this study shows that the improved CVaR-based Portfolio Optimization Method is able to provide a better optimization of a diversified bond portfolio in terms of reducing the expected loss, and hence maximizes the returns.