INCORPORATION OF CONDITIONAL VALUE-AT-RISK INTO MEAN-VARIANCE OPTIMIZATION FOR PORTFOLIOS OF HEDGE FUNDS

In this paper we aim to search for a systematic optimization model that can properlymeasure hedge fund risks and can optimize capital across Canadian hedge fund portfolios thatcan cater to investors’ risk appetites. As the characteristics of hedge funds returns imposedifferent layers of risk from traditional equity and bond investments, the conventional meanvarianceoptimization would not accurately capture the risk associated with non-normaldistributions and negative skewness. The process requires a different approach that modifiesthe drawback of a mean-variance optimization to take non-normal and asymmetricdistributions into consideration. The research of this process leads to a Mean-ConditionalValue-at-Risk (CVaR) optimization. CVaR measures the mean expected short fall betweenvalue-at-risk and excess losses that reflect the risks of kurtosis and negative skewness.Combining the cluster analysis to overcome variation of correlation issue and Mean-CVaRoptimization, we found the Mean-CVaR optimization model that will serve the requirementsof guiding investors’ capital allocation among hedge fund strategies.