FT (Nov 5, 2001, pg. 22): "Study questions risk profile of hedge funds"
COMPANIES & FINANCE INTERNATIONAL: Study questions risk profile of hedge funds
Financial Times, Nov 5, 2001
By PHILIP COGGAN
The merits of hedge funds may have been overstated because of faulty statistical analysis, according to a study of their performance over a six-year period.
While funds appear to offer attractive average returns, investors have a much greater risk of suffering a heavily negative outcome than with conventionally managed funds.
The classic example of this tendency was Long-Term Capital Management, the US hedge fund, which delivered high returns through the mid-1990s but blew up after the Russian debt crisis of 1998.
Hedge funds are alternative investment vehicles that follow different strategies from those used by traditional fund managers - for example, they may focus on the debt of distressed companies or take bets that certain share prices will fall while others will rise (so-called long-short funds).
Those marketing hedge funds claim they focus on absolute (rather than relative) terms and deliver higher returns, for a given level of risk, than are achieved by traditional fund managers.
But a study by Chris Brooks and Harry Kat of the University of Reading* says that hedge fund returns do not follow the normal distribution or "bell curve" pattern. Instead they demonstrate statistical properties known as negative skewness and kurtosis.
In short, there are more extreme outcomes than should occur under a normal distribution and more of those outcomes are negative than should be the case.
The academics, who based their study on figures compiled from hedge fund indices over 1995-2001, also find that most hedge fund categories show surprisingly high correlations with the US stock market, despite the unorthodox strategies they claim to follow.
And some classes of hedge fund, notably in the convertible arbitrage and distressed securities indices, show high autocorrelation - the return in a month tends to be very similar to that of the previous month. This probably indicates the difficulty in finding market prices for the illiquid securities in which they specialise. But it may also mean returns from such funds appear far more stable than they actually are.
All told, the academics conclude that investors using traditional statistical approaches such as the mean and standard deviation, will allocate too high a proportion of their portfolios to hedge funds.
*The Statistical Properties of Hedge Funds and their Implications for Investors, working paper, October 31 2001
Copyright: The Financial Times Limited
____________
Note: I culdn't find the working paper, but here's some other WPs: isma.rdg.ac.uk. |
