“We’re seeing a growing number of hedge fund managers look at stress testing, in addition to VaR [value at risk], as part of their risk management,” says Lance Smith, CEO of Imagine Software, a software developer for investment management solutions. “The limitations of VaR have become more obvious in the light of some extreme market events such as [hedge fund] Amaranth’s blow-up.”
Smith continues: “VaR looks at historical correlation of risk factors, which is essentially a type of average. But in calculating risk in the event of an extreme situation, stress testing is more appropriate. VaR is expressed as a multiple of a typical daily movement. You can estimate the daily volatility of returns and then, if you see frequent P&L swings greater than that, you know there is something awry in the portfolio. It’s a good calculation of likely events. But in low probability and extreme situations, correlations are very different and stress testing against different criteria can help show what level of risk a portfolio has.”
Geoff Allan, director of prime services at Credit Suisse in London, points out other limitations of VaR. “VaR is not appropriate for illiquid securities, such as emerging market equities or small-cap stocks that lack trading volumes. VaR either assumes a normal distribution of stock returns or requires an adequate supply of historical market data, so for illiquid instruments it is not an appropriate indicator. Also, it is unsuited to any situation where historical price is not a good indicator of future risk. For example, in the case of merger arbitrage transactions, as a rule, following the announcement, the acquiring company will experience a drop in share price and the target will see an increase. Until the corporate action completes, there is a risk that if the deal might break up and stock prices return to around previous levels, this type of price move is not something that could be predicted using VaR and historical volatility.”
Mark Friedman is principal of AM Investment Partners, which runs three hedge funds. “I’ve always been a big believer that stress testing is more relevant than VaR in risk management,” he says. “VaR tends to be favoured by banks as the information flows up a bigger chain of command, so having just one number makes it easier for them. But stress testing allows us to see a variety of different risks at once and is therefore a more useful guide to risk.
“We use a variety of different stress tests and can shock our portfolios for virtually anything, depending on the strategy. That is, equity moves, movements in volatility, and movements in credit spreads, interest rates and borrowing costs.” The firm runs stress tests daily, weekly and monthly.
AM Investment Partners does still provide VaR to some clients but Friedman believes there will be a shift to stress testing. “Some clients prefer to see a VaR number as they are more familiar with VaR and have used it as a risk measurement in the past but we are seeing more clients gravitate towards ‘what if’ scenarios.”
Prime brokers themselves are increasingly using stress testing in their determination of margin requirements.
“The margin requirements of a prime broker directly affect the leverage its clients can achieve, which in turn directly affects the client fund’s returns,” says Smith. “So, on the one hand, the competitive landscape of prime brokerage puts pressure to reduce the margin calculation, while still maintaining prudent margin requirements. Rule-based margin is often onerous in this respect; it may be safe, but it may not be competitive. In particular, rules-based margin does not give sufficient relief across related products and strategies.”
Credit Suisse’s Allan says VaR has its place but should be used in conjunction with stress testing, because of its limitations for gauging risk.
“VaR is a decent tool, but stress testing should be used to complement it, and stress testing is becoming more popular as technology improves. We stress test every day at the bank, whereas we used to do this perhaps every couple of weeks or once a month. We run our portfolios through historical scenarios such as Black Monday or 9/11, and through ‘what if’ scenarios so we can work out whether we have the appropriate level of margin to cover the risk, and understand in what situations our portfolio will lose money. Not all prime brokers stress test the book the way we do at Credit Suisse, many are still coming up the curve technology-wise. We’ve been running Imagine now for nearly four years, meaning our systems are fine-tuned and well connected.”
Imagine’s president, Steven Harrison, agrees that technology has been lacking in allowing managers and prime brokers to be able to stress test. The firm has developed a system, however, that enables managers and prime brokers to run customized multi-factor stress tests through an online product.
This move to stress testing by prime brokers is further encouraging hedge funds to follow suit. “Hedge funds that have the capability have begun to include the same stress tests on their positions so that they can determine what their margin is likely to be, as well as to explore strategies to reduce the margin requirement,” says Imagine’s Smith. “The smaller the hedge fund, the more urgent the need for stress testing. This is because, unlike the mega-funds, small funds are not well diversified across strategies and market segments, so that the assumptions of normality underlying parametric VaR are not as well approximated.”
Allan says: “Hedge fund managers are looking at risk now more than ever before, a significant driver of this is the growth in the institutional investor base. This type of client has a structured investment process: prior to allocating capital, questions are directed to fund managers about how they measure risk, what systems and processes are in place, and content of the monthly risk-reporting pack.”
