Existing measures of risk just don't hack it when it comes to hedge funds.
The problem is many hedge fund managers are loath to disclose information that would reveal their positions, while many institutional investors insist they understand the risks they are taking on before investing in the vehicles.
"Right now, it's a Mexican standoff between the hedge funds and the pension funds," said Leo de Bever, senior vice president-research and economics of the C$68 billion (US$43 billion) Ontario Teachers' Pension Plan Board, Toronto.
Normally, one would think pension funds would be calling the shots. But current demand for hedge funds outstrips supply by a factor of more than two or three, he said. "That puts (pension) funds in the weird position of saying, `If I buy into this, I don't know what risk I am taking, and I have troubles with that,"' Mr. de Bever said.
What the industry needs are risk analytics that provide investors with a snapshot of a hedge fund's risk exposures without revealing its underlying investments, argues Andrew W. Lo, a well-known Massachusetts Institute of Technology finance professor, in a paper to be published in the Financial Analysts Journal's November/December issue.
Efforts under way
Some efforts already are under way. Tremont Advisers Inc., Rye, N.Y., and RiskMetrics Group, New York, recently teamed up to develop a web-based system that provides such analytics. Peter Bernard, executive director of RiskMetrics, hopes that low fees charged to hedge fund managers, pressures from large investors and the threat of regulatory action will induce hedge funds to participate. He hopes to start issuing reports within three months.
And the Investor Risk Committee, a group comprising institutional investors and hedge fund managers that is sponsored by the International Association of Financial Engineers, is developing disclosure standards based on risk measures.
Others, such as Capital Market Risk Advisors Inc., New York, are serving as intermediaries between hedge funds and investors, evaluating risk exposures without giving proprietary information. "What I believe is when (investors) say they want transparency, what they really want is some sort of predigested analysis that highlights the risks, the changes (in exposures)," said Leslie Rahl, president.
The problem, she added, is that managers calculate risk levels differently. There are seven or eight ways of calculating value at risk, a widely used risk measure. And she doubts there ever will be an industrywide standard, noting that banks, which are highly regulated, have failed to coalesce around a single standard.
In his paper, Mr. Lo is critical of VAR, which provides an estimate of how much an investor can lose in a defined time period. VAR, for example, cannot capture the wide range of risks that hedge funds exhibit, such as style, fundamental analysis, short squeezes, yield-curve models, optionality, credit risk and macroeconomic factors, wrote Mr. Lo, who is on leave as the Harris & Harris Group Professor at the MIT Sloan School of Management. He also serves as chief scientific officer of Alpha Simplex Group LLC, a Cambridge, Mass., money manager.
Also, VAR might not be suited to certain types of investments, such as emerging market debt, risk arbitrage or convertible bond arbitrage, Mr. Lo wrote. In addition, "VAR is notoriously difficult to estimate," he wrote.
Other experts agree some versions of VAR are inadequate but think more sophisticated uses of the tool are more reliable. Kelsey Biggers, vice chairman of Measurisk LLC, New York, said using VAR on historical returns of an entire portfolio won't show much about future risks. His firm conducts Monte Carlo simulations on the security level, which he says provides a more reliable measure of sensitivity of individual securities to different factors. What's more, VAR needs to be combined with stress testing, he said.
Echoed Mr. Bernard: "We certainly would agree that VAR as a single statistic is far from sufficient in giving one a (measure of the) total risk of the hedge fund."
There are other problems with applying other traditional measures of risk, Mr. Lo wrote:
Survivorship bias in hedge-fund databases creates problems for "the unwary investor seeking to construct an optimal portfolio of hedge funds."
Traditional measures, such as mean-variance analysis, fail to capture the risk of dynamic trading strategies, particularly missing the risk of infrequent but dramatic drops in value.
While many hedge fund strategies appear to be relatively uncorrelated with market indexes, "phase-locking" behavior - such as occurred during the Russian debt crisis in summer 1998 - and other "non-linearities" can radically alter results and are not picked up in traditional measures.
Creditor relationships and liquidity risks often do not hit the radar screen.
Incentive fees for hedge funds managers, providing enormous upsides and limited downsides, "can induce excessive risk-taking behavior if ... not properly managed." Risk preferences of both investors and managers should be considered.
The root of the problem is that hedge fund managers and institutional investors look at the world very differently, Mr. Lo wrote. Hedge fund managers typically have very broad discretion in making investment decisions; their trading strategies are highly proprietary; returns in most cases are paramount; regulatory constraints and compliance rules generally drag on performance.
In contrast, institutional investors typically need to fully understand an investment process before committing to it; they must understand a manager's risk exposure; and they measure performance not only by returns, but also by risk. Institutions are highly regulated and prefer structured, investment processes, he said.