PT - JOURNAL ARTICLE AU - Clifford De Souza AU - Suleyman Gokcan TI - Allocation Methodologies and Customizing Hedge Fund Multi-Manager Multi-Strategy Products AID - 10.3905/jai.2004.391060 DP - 2004 Mar 31 TA - The Journal of Alternative Investments PG - 7--22 VI - 6 IP - 4 4099 - https://pm-research.com/content/6/4/7.short 4100 - https://pm-research.com/content/6/4/7.full AB - This article identifies many of the issues crucial to hedge fund investing that we believe may be ignored by the uncritical investor. We believe that some of these issues can be taken into account by a systematic investment methodology and outline a method for doing so. We lay out a complete methodology for a hedge fund strategy allocation. Based on the monthly return data over the period 1990–2002 we find that the return distributions of most hedge fund strategy indices display significant negative skew, as well as serial correlation and unstable correlation structures. These findings have important consequences for assembling portfolios of hedge funds. We demonstrate that although hedge fund strategy indices are highly attractive when only mean and variance are considered, this is much less the case for some strategies when the statistical effects of serial correlation are considered. We optimize with respect to conditional value-at-risk to take the negative skew into account. When we compare our results with mean-variance optimization, it becomes clear that mean-variance optimization underspecifies the risk in hedge fund strategies. This article isolates some of the historical properties of hedge funds and makes, we hope, a strong case for their systematic inclusion in a modern portfolio.