When considering any investment, prospective investors and their advisors look first at its returns and then at the risk patterns of its past. For mutual funds and traditional equity investments, these past return performance patterns are likely to be more indicative of what may be expected in the future but, when it comes to hedge funds, this is not necessarily the case.
Mutual funds are subject to many more restrictions as to the strategies and investment vehicles they employ. For example, with limited exceptions, they can only buy instruments whose value appreciates when the value of the underlying asset increases and therefore the attributes of their risks and returns are very similar to those of the market they invest in.
Not so for hedge funds. There is a lot more to hedge fund investing than looking at average performance whether one considers risk, return or any other of the more or less exotic ratios used in this industry. It is actually the very reason why hedge funds utilise so many different measures and ratios not found in traditional investment settings. These yardsticks are all designed to help advisors and investors alike to have a better handle on analysing hedge funds and many of them are pretty useful up to a point. They are still, however, far from being able to offer a complete picture or provide a true indication as to how any particular hedge fund will do in the near future.
Let's consider the two imaginary, local hedge fund managers whose results are shown in the table below. HedgeGlow likes to make small gains but when he loses, he loses big time. HedgeBlow (who should have stayed away from money management) likes to try his luck with a strategy by which his daily gains or losses are the same but he is more likely to lose than gain on any single day. Undoubtedly HedgeBlow pins his hopes on an improbable home run that would see him with more oneday gains than oneday losses, which is contrary to what his strategy profile would suggest. Now let's look at the expected rate of return and the risk characteristics of these two hedge funds.
HEDGEGLOW  HEDGEBLOW  

Daily Return  +0.2%  Probability 
Daily Return  +0.5%  Probability 
Assume that there are 250 business days in each year and consider the performance of these two managers over a four year period, that's 1,000 business days. Let's suppose that the following returns behaviour was observable during this period, exactly as would be expected from the probability chart above.
HEDGEGLOW  HEDGEBLOW  

No. of positive days  999  400 
No. of negative days  1  600 
Annualised rate of return  22.36%  22.36% 
Risk (annualised standard deviation)  47.6%  7.75% 
Based on the fouryear performance history, the two managers
have realised identical returns, albeit with different risk
characteristics. This is exactly what was expected given their
respective risk and return profiles.
But what happens when the outcomes are not exactly as expected from each manager's return probabilities? Let's look at the following scenario:
HEDGEGLOW  HEDGEBLOW  

No. of positive days  1,000  450 
No. of negative days  0  550 
Annualised rate of return  64.5%  12.03% 
Risk (annualised standard deviation)  0%  7.87% 
There is a huge difference  for HedgeGlow especially  between the true characteristics of these funds and what one might infer from their actual performance. Looking only at past risk and returns, investors would likely line up to put their money into this new wonderfund called HedgeGlow. But why? Only because their perception of the fund is very different from its true behaviour.
What I am trying to show here is the following:

Managers can, and sometimes do, use strategies whereby losses are unlikely but when those losses occur they are large. This kind of hedge fund behaviour is very misleading. It gives an artificial impression of the fund being a low risk investment and its true behaviour only emerges relatively rarely. The fact that such a large loss may sometimes not have arisen over a long period can give investors a false sense that the fund is less risky than it really is.

The past returns of a hedge fund reveal only one set of possibilities from the many potential results given that fund's specific risk and return characteristics.

Hedge funds change and over time they may pursue different strategies with significantly different risk and return characteristics. Calculating average performance for the whole period in which a number of different strategies have been used may provide a greatly distorted picture.
The problem is that the return patterns of hedge funds can be a lot more varied than those of mutual funds, due to their ability to pursue a wide variety of strategies using an equally wide variety of instruments. Mutual fund investments are limited to being long on investment vehicles like stocks with clear limitations on maximum exposure and market caps. This means that their return patterns will be a lot less varied and a lot more dependent on the market they invest in than those of hedge funds.
These are the factors that make the analysis of hedge fund returns more of an art than a science. The main issues to be taken into account when considering hedge fund return patterns are:

Strategies pursued by hedge funds are perfected, discarded and/or replaced by their managers over time. Grouping returns together for periods in which completely different strategies may have been used can make the analysis of a hedge fund significantly less reliable.

Assets under management for hedge funds change significantly over time, going from a small pool to medium size to large assets or decreasing significantly from large to mid size. And remember that these terms are all relative; large means different things for different hedge fund strategies and managers. Performance behaviour during a change and again after the change has taken place, may vary quite a bit.

Exceptional performance  good or bad  will cause a manager to act slightly differently than normal. Performance patterns may be markedly different due to desperation and the need to achieve a good result no matter the cost, or to a belief in invincibility after a long streak of positive, higher than normal results.

Past returns, month by month, and the patterns they create, are not indicative of the returns and patterns which would prevail under certain market conditions.
After identifying the limits of hedge fund analysis, the questions are 1) whether to discard, totally or in part, this way of analysing hedge funds quantitatively; and 2) is there a better method of analysis offering a more accurate sense of their risk and return expectations? To answer these questions certainly requires a great deal more scrutiny and indepth research. I believe however, that a serious "analyst", by which I mean an investor, advisor or professional analyst, should make the following modifications when addressing hedge funds:

Test whether returns data apply to distinct periods. If so, treat each period as a separate unit and do not base conclusions on an analysis only of the period as a whole.

Be aware of the probabilities of returns under different market conditions.

Perform a sufficiently large number of MonteCarlo simulations based on the returns and their respective probabilities and then analyse the best and worst cases.

Predict risk of the hedge fund based on the probabilities and return patterns and compare these to risk levels of the MonteCarlo simulation risk levels.
The investment industry badly needs some different ways to analyse hedge funds rather than just looking at them through the same tools as those used for traditional investments. This process is likely to evolve over time but it will always exhibit less predictive power and more subjectivity than we are used to in the traditional investment industry.
Miklos Nagy, CFA, CFP is a hedge fund analyst, Chairman of Canadian Hedge Watch Inc and Chief Executive of Quadrexx Asset Management Inc, Toronto, Canada.