Hedge funds are a mystery to many people. They can be confusing and investors are wise to seek the advice before investing. MarketWatch posed the question of whether some people are statistically better at selecting hedge funds and predicting hedge fund performance in general. Two teams of advisors were considered – the first, a group of professional investment advisors running a “fund” of hedge funds and the second, a group of monkeys.
Statistical Proof
What actually happened was that a group of researchers in Europe looked at “a broad sample of 1,300 funds-of-funds from 1994 through 2009, and analyzed just how much value they actually created.” The team included Benoit Dewaele and Hugues Pirotte of the Universite Libre de Bruxelles (the Brussels Free University in Belgium), and Nils Tuchschmid and Erik Wallerstein of the Geneva School of Business Administration in Switzerland. The team published its results in a paper called, “Assessing the Performance of Funds of Hedge Funds.” It found that “just one fund in 20 actually added risk-adjusted returns above those of the underlying hedge-fund indices.”
It Gets Worse
Perhaps worse is the fact that not only did so few outperform the market, but so many did not even produce results on par with hedge fund indices. “Nearly half of all Fund of Hedge Fund managers, the academics report, delivered ‘negative after-fees alpha when benchmarked against the hedge-fund indices.’ In other words they couldn’t even keep up with the index,” suggesting that average hedge fund picking skills amongst the sample selected were “close to non-existent in the first place.” Even when the researchers compared the performance of the “funds of funds” with the randomly selected stocks, the differences were minimal and amongst those differences the fees associated with the funds of funds wiped out any real benefit.
It may seem hard to believe, but here is a graph that proves the point.
Pictured is a graphical interpretation of “Cross-sectional distribution of mean returns of real and simulated FoHFs (1/1995-8/2009). 1315 drawings of 15 hedge funds (among those available) are performed to produce a sample of 1‘315 simulated FoHFs to be compared to the sample of 1‘315 real FoHFs. Periods of existence and fees are perfectly matched with those of the real FoHFs. The figure presents the distribution of mean returns over the entire corresponding period for each FoHF (with varying lifetimes), either simulated or real. It is the direct distribution of FoHFs‘ returns, without any averaging into a portfolio as in Figure 1. The present figure shows the exercise with a 12-month reallocation frequency (the random choice is performed at the beginning and then every year for the same period of existence as for the real hedge fund, with a verification still for the hedge fund that would have been liquidated in the meantime). Returns are computed on a monthly basis. Extreme mean returns are mostly generated for FoHF with a very short existence, reduced by the mentioned corrections of selection biases.”
