Riskdata Uses Factor Data to Highlight Time Bombs |
Date: Friday, May 16, 2008
Author: Bill McIntosh, Senior Financial Correspondent, Hedgeworld.com
LONDON (HedgeWorld.com)—The high-profile effect of the subprime crisis on alternative investments, notably credit strategies such as fixed income and distressed investing, is symbolized by the large number of blowups among managers whose track records prior to the summer of 2007 didn't suggest potential high risk. However, in a new study, "Factor Based Vs. Return Based Models," Riskdata concluded that assessing risk through return-based models using Value at Risk or volatility is insufficient. Instead, the factor-based model explores the relationship between funds' performance and a wide variety of different market factors ranging from equity indexes and interest rates to commodity prices or credit spreads. "The problem of the subprime crisis is one of risk transparency," Olivier Le Marois, chief executive of Riskdata, said in an interview. "If you are invested in a fund that is meant to be unrisky and suddenly it falls 40%—as an investor, you need a way to know that this is possible. It is a question of leverage and factors. And that can be tracked by linking the performance with the factor." In an analysis of 3,200 hedge funds, Riskdata found 729 funds that experienced large drawdowns not only compared to their volatility (more than 2.3 times) but also compared to prior maximum drawdowns (more than 2 times past drawdown). The perimeters used meant that there was nothing in the track record of these funds that could have alerted an investor to the possibility of such a high level of losses. The issue Riskdata wanted to explore was whether it was possible to detect the time bombs that these funds contained. Overall, the funds that experienced high drawdowns fell an average of 9.4% from June 2007 to March 2008. (And all the funds in the survey had a track record stretching back to at least December 2004.) Using a return-based risk model, an investor who in June would have only accepted to invest in hedge funds exhibiting "normal" risk patterns would have slightly over-performed during the crisis compared to an ISO-weighted benchmark. The result is brought about by the elimination of extreme risk takers, but the benefit of this elimination is somewhat offset by the fact that it also eliminates successful risk takers, and fails to detect time bombs. However, using a factor-based model incorporating a particular market variable or variables, an investor comparing drawdowns with predicted ones (using the factor model) would have outperformed the benchmark by 4%. Crucially, this was achieved through a reduction of time bombs, while keeping successful risk takers in the portfolio. "The study shows that with an appropriate risk system you can make risk transparent," Mr. Le Marois said. "You can then decide how to invest on that basis." In a hypothetical example, a credit manager's returns might be explained by the narrowing spread between government and investment-grade bonds over the period 2003 to 2007. Spice up with leverage and the returns would have been highly desirable with a low-volatility profile. But once volatility in the spread shifted back to the levels seen in 2001 and 2002, the return profile became very different. "You want to be able to discriminate between the manager that is delivering alpha and the one who benefits simply from the bubble," Mr. Le Marois said. "That is, discriminate between a passive beneficiary and an active one. To do that you need a risk system that look at the drivers of the returns rather than the returns themselves."
Reproduction in whole or in part without permission is prohibited.
