Alternative investments have become the fashionable asset class of choice for pension funds, charitable endowments and wealthy individuals. The initial attraction was the outsized returns of the 1990s, but the concept of being able to show positive performance in both favourable and unfavourable markets has proved equally compelling. Many hedge funds promise less volatility than long-only investors, and this is appealing to fiduciaries. After the collapse of Long-Term Capital Management, blow-up risk became a selection factor for institutions. Nobody wanted to lose significant capital with a single manager, especially one who promised a “market neutral” approach.

Hedge fund managers began to talk about how few months they had with negative returns, and somehow low volatility became an objective, not a by-product. In 2003 when the Standard & Poor’s 500 was up close to 30%, the average hedge fund only had a net return of half that, and their managers defended themselves by saying they achieved that return with lower volatility than their long-only competitors. I began to feel strongly about this after some conversations with my old friend Dick Elden, founder of the first US fund-of-funds, Grosvenor Capital Management, in 1971.

In my view this focus on low volatility is excessive. Very few institutions or individuals use only one hedge fund, where concern about volatility would be understandable. Most institutions use at least five, with many having a dozen or more. This is consistent with my view that investors should have a portfolio approach to hedge fund investing, using managers with different approaches and adding some and eliminating others each year. Within certain limits, plan sponsors should pick managers with strong performance and worry less about volatility. As Warren Buffett said, “I’d rather have a lumpy 15% return than a smooth 12%.”

Although many investors in hedge funds have embraced the portfolio concept, they seem to put low volatility at a premium when selecting individual managers. I believe this is a mistake. As long as they pick a group of managers with different investment approaches so they are not closely correlated, they would be better served by choosing managers with superior long-term returns, even if they have achieved those returns with relatively high volatility. When the individual high volatility of a number of non-correlated high-performance managers is combined in a single portfolio, the overall return will often be higher and the volatility will be lower than with a group of managers who were selected because each had reasonable returns with low volatility. The high volatilities of the individual non-correlated managers will offset one another, producing higher returns with lower volatility for the overall portfolio.

I was intuitively convinced this idea was right, but I knew none of you would take my word for it. I sought out Michael Peskin and Bryan Boudreau, who run Morgan Stanley’s Asset Liability Strategies group in Global Capital Markets, and asked them to develop a model to prove (or disprove) the concept. Their results were encouraging, but they had some caveats. Exhibit 1 shows two portfolios of 25 hedge funds. In the first one, each individual hedge fund has an expected return of 7%, a standard deviation of 6%, a correlation of 40%, a risk-free rate of 4%, and a Sharpe ratio of 0.5. In the second group of 25 funds, each has an expected return of 10%, a standard deviation of 15% (2.5 times the first group), a correlation of 20%, and a Sharpe ratio of 0.4, modestly lower than the less volatile group of funds. According to their analysis, the expected portfolio return of the first group of funds is of course 7%, the portfolio standard deviation is 3.91%, and the portfolio Sharpe ratio is 0.77. In the second group of funds the expected return is 10%, the standard deviation of the portfolio is 5.53%, but the Sharpe ratio is higher at 1.08. As Peskin and Boudreau point out, “The latter portfolio clearly dominates (a much higher portfolio Sharpe ratio) even though the individual funds have considerably higher volatilities and lower Sharpe ratios.”

They go on, “The proper performance objective of funds of funds is to maximize portfolio Sharpe ratios. This means paying more attention to low cross correlations. This also requires finding those hedge funds that provide alpha idiosyncratically, are willing to swing for the fences, and rely on the fund of funds approach to diversify their high idiosyncratic risk.”

The demand for hedge funds has picked up in a low-return environment, and hedge funds have been able to increase their fees (from about 1% of assets and 20% of the appreciation to about 1.5% and 20%). The higher management fee places a burden on “net” performance in a low-return environment. It also makes the management fee the basis of a very lucrative business, rendering the incentive fee less important. Too many hedge funds have become asset gatherers, reducing risk at the individual fund level and ignoring the reality that investors can get the right risk characteristics from pooling.

It is not clear, however, that the higher-return, higher-volatility hedge funds will have the necessary lower correlation for pooling to work. It will take some serious effort by an institution to put together a group of high-return, high-volatility, non-correlated managers employing different strategies. Simply having a group of long/short equity managers focusing on different sectors or capitalisation sizes won’t work.

An endowment or retirement portfolio does not consist of hedge funds alone. The correlation of the “alternative asset” portfolio must be considered in conjunction with the long-only portion of the asset mix. Our Marty Leibowitz has found that the performance of most long/short and long-only managers is highly correlated with the US market regardless of where the manager is invested. In addition, the net equity exposure of many hedge funds is similar, so the idiosyncratic risk of the portfolios is critical.

While plan sponsors may benefit from seeking individual funds with higher returns and resultant high volatility, the fund managers themselves may be wary of such an approach. Most hedge fund managers have a significant portion of their personal assets invested in their own funds and little or nothing in other hedge funds. Because they “eat their own cooking” they may not want to endure a big down year, even if it follows and/or precedes years of exceptional appreciation. Since hedge fund managers know they must maintain a critical mass of funds under management in order to earn a management fee sufficient to support their infrastructure and adequately pay their analysts and portfolio managers, they may not want to risk the loss of accounts and the difficulty of attracting new business that comes with a higher level of volatility. These factors encourage hedge fund managers to take a relatively risk-averse, low-volatility approach in their portfolios.

Hedge fund risk aversion was partly behind the mediocre performance of most long/short equity funds in 2003. There is some evidence that this has continued through this year. International Strategy and Investment (ISI) maintains a survey of the net exposure of hedge funds. In early 2003, before United States troops entered Iraq, the funds in their sample were only about 35% net long. As the market continued to rise during last year, the net exposure rose to about 60% in December, where it peaked. For most of this year the net invested position has been below 50%, but over the last few weeks, as the market has done better, exposure has risen sharply from about 45% to 52%. A chart of hedge fund exposure looks very much like a chart of sentiment (the Ned Davis Crowd Sentiment Poll, for example). These measures of investor enthusiasm tend to illustrate pessimism when opportunity is developing and optimism when prices are full, proving that most hedge fund managers are ordinary people and not “masters of the universe” after all. This, also, may help explain why many hedge funds are failing to keep up with the indexes.

In my mind to achieve truly superior performance a hedge fund manager must identify concepts and themes that are undervalued and have considerable long-term potential. They can do some trading around their core positions, but the thrust of their performance will come from having a substantial weighting in a number of non-consensus ideas that turn out to be right. This often requires both conviction and patience, but along with those manager qualities comes portfolio volatility. Both managers and their clients have to recognise that above-average volatility may be a characteristic of superior investment performance. If they are not willing to tolerate that volatility, they may be saddled with mediocre performance in a low-return environment or worse. Then they will have to think hard about whether their aversion to risk is worth a fee of 1.5% and 20% of the appreciation.