Since its introduction, the mean-variance methodology became the primary tool for portfolio diversification used by the majority of pension and mutual funds globally. However, despite its popularity, the mean-variance approach suffers several important drawbacks limiting its applicability, especially, when it comes to hedge funds.
First, the estimate of risk by variance is only appropriate, when returns are normally distributed or investors exhibit quadratic preferences. Examination of returns of different asset classes shows that traditional instruments, like stocks or bonds, demonstrate distribution more or less approximated by normal, whereas derivatives evidence a high level of irregularities as skewness and kurtosis excess. Numerous studies evidence that most of hedge funds expose asymmetrical returns, therefore, making the mean-variance model hardly applicable in principle.
Second, the mean-variance framework assumes that investors focus on a single time horizon and will never alter their asset allocation once it is chosen.
Third, according to the mean-variance approach, the main objective of investors is to minimize the volatility under the defined mean of returns or vice versa. It does not cope with today’s sophisticated investment techniques and instruments. For example, the objective of a fund of funds may be in minimization of fund’s correlation with the chosen index, thus implying a market neutral strategy.