October is likely to prove a terrible month for hedge fund managers across asset classes because of sudden changes in the way various risks correlate, says William Reeves, principal at UK-based BlueCrest Capital Management, an alternative investment platform partly owned by Man Group. Reeves manages the firm's flagship relative-value fixed income fund, a low-return, low-volatility product popular with many institutional investors, and believes the Street's misunderstanding of correlation risk is setting investors up for losses.

Reeves does not set out return targets, but volatility targets, with the flagship fund aiming for a low, 4% annualized volatility target by making lots of niche bets across the bond universe. BlueCrest also offers higher-octane products.

Given this framework, Reeves believes the biggest risk is market perceptions of risk, and a reliance upon Value-at-Risk (Var) as a measure of risk. Var attempts to give financiers a probability of a portfolio's current risk, not just a historical projection, and while it has its uses, Reeves thinks too many financial institutions are relying on Var as a magic bullet, a single number that can describe a portfolio's total risk.

"But Var depends on correlations," Reeves says. "Correlation risk is hard to assess because it's always a moving target, unlike, say, duration or credit risk."

Nonetheless he believes it is correlation risk that does damage to hedge fund managers. The downgrade of GM and Ford debt to junk status in May served as a catalyst to losses, but the actual damage happened because the correlation of risks among various asset classes changed unexpectedly.

The real problem for managers in May was the collapse of the price of credit insurance, expressed in the form of credits on implied volatility, thanks in part because of retail investor behaviour. For the past two or three years, retail has snapped up structured products such as guaranteed funds issued by investment banks and fund companies.

Retail investors are, essentially, selling volatility. They are not checking their Bloomberg screens every hour; they review their fund performance once or twice a year. So they are insensitive to volatility movements, which financial institutions may find very destabilizing - such as the auto downgrades in May.

"There was a huge mismatch between risky credit exposures and where the Street was trying to position itself," Reeves explains. "The Street assumed retail investors would stop selling credit risk, so it bought insurance. But investors didn't sell credit risk."

In 1998, it was hot money such as LTCM which was leveraged. Today it is the retail market that is leveraged. But retail investors do not behave like financial institutions so when GM and Ford got downgraded, retail did not panic and sell. Professional investors did, however, and got burned.

"Fixed income investors are now trying to anticipate when retail sells, which I think is a mug's game," Reeves says. "Credit funds lost money when they got bearish too soon."

This episode reveals that reliance upon simply calculations such as Var means financial institutions think they understand their positions' risk/return profile and think they can neutralize that through hedges, but are vulnerable to shifts in correlation risk. "The market is complacent and anyone running spread-related portfolios needs to actively manage correlations," Reeves says.

He thinks this lesson may be driven home over the course of October. "This is going to be a horrible month for hedge funds, worse than May," he predicts.

Three factors are conspiring to create losses across asset classes. First, markets have misjudged central bank policy. The US Federal Reserve rose rates despite the damage of Hurricane Katrina, the Bank of Japan is making hawkish noises about withdrawing liquidity, and the European Central Bank recently switched its rhetoric from neutral toward a bias toward future monetary tightening. These actions, particularly in Europe, surprised the market and has led to sudden liquidation of positions at the front end of money market funds.

The second event was the sudden collapse of Refco in the US, which has raised the spectre of counterparty risk in an environment of shrinking liquidity. Although this combination does not pose a systemic risk, it has led to sell offs at the front end of the yield curve. "It's not a good environment for paper assets," Reeves says.

Finally, hedge funds have overall had a good year, particularly in equities, and managers are booking profits in October. Many institutions are keen to avoid the market dips common in December and the "year end" has shifted earlier in the year.

The result of these three events: "There's been a huge jump in correlations between asset classes in the past two weeks," says Reeves. "This general liquidation has not spread to retail, however; most retail investors have no idea how Refco impacts their portfolio."

Reeves says for hedge funds with low risk exposures, these spikes in correlation volatility can present great opportunities to structure market-neutral trades. "People are selling what they've got, not because they want to. Var becomes a problem because whatever your risk system says you're running, you're running more. It underestimates the volatility you'll see in your P&L. So wobbles in correlation can lead to liquidation trades."

This situation is likely to last as long as retail continues to damage professional investors who must mark to market. Eventually the investment banks will package new structured products that will convince retail to buy volatility (ie spread insurance) instead of sell it. "When this happens it will provide us with a great opportunity, but for now we're sitting on our hands," Reeves says.