Thought Winter 2008

RISK: You Can’t Manage What You Can’t Measure

E-mail

  Andrew Lapkin
  President
  Measurisk, LLC

Risk measurement is critical for effective portfolio management, yet more than ever, many traditional risk measures are proving to be problematic. Institutional investors and consultants often use historical measures exclusively when they should also be using forward-looking tools to complement their current approach. The most common historical measures include standard deviation and the Sharpe Ratio, while forward-looking tools include value-at-risk (VaR), sensitivity analysis and stress testing. Effective risk measurement requires a portfolio of tools that rely on current positions, historical data and modeling tools.

Holdings-Based Risk Assessment
VaR is a forward-looking measurement of risk, an assessment of potential losses a portfolio may experience in the future. Standard deviation, beta and tracking error are all measures of what has occurred in the past. Mathematically speaking, all of the measures are rooted in many of the same principles—the key difference being that VaR is calculated using the portfolio’s current holdings and current market conditions, while more traditional measures rely solely on historical returns. Since portfolios and markets change, VaR is a more effective way to understand a portfolio’s risk than historical measures. Historical measures can still be helpful, but they should not be the only methods used in assessing current and future risk in a portfolio.

Importantly, VaR is just the starting point, as there are a number of closely related measures that should be used to understand the risk profile of an investment manager.

Relative Risk: Risk versus a Benchmark

The focus of risk measurement is often performance against an asset-based benchmark (e.g., an index) or liability-based benchmark (e.g., a pension benefit liability). Tracking error is the most popular of these measures, but it captures only the volatility in relative performance against the benchmark, rather than the absolute amount by which a manager may exceed the benchmark in the future. As an extreme example, two managers who separately outperformed the benchmark by 1% and 10% every month for a year would each have a one-year tracking error of zero! As such, a manager’s tracking error may not be a good indicator of future performance. In order to estimate that, one must use relative VaR.

RISK: You can't manage what you can't measure

Relative VaR
Relative VaR (R-VaR) is calculated by modeling the differences between the portfolio and the benchmark. Assume a benchmark is composed of $20 of asset A and $15 of asset B, while the manager holds $25 of asset A and $10 of asset B. The risk of underperformance to the manager would be in scenarios where the value of asset A goes down (the manager holds more of it) or asset B goes up (the manager holds less of it). R-VaR is thus calculated using these over and under weights—in this case, the risk of holding an extra $5 of asset A and being short, or holding $5 too little, of asset B. The resulting risk number thus represents a forward-looking tracking error and provides an accurate measure of how much a manager may underperform the benchmark in the coming month.

Marginal VaR
Another important consideration for institutional investors is diversification across managers. Marginal VaR (M-VaR) measures the risk contribution of each manager to the total portfolio VaR. M-VaR of a given fund is calculated as the difference between the VaR of the full portfolio and the VaR of the portfolio after removing the individual fund. The lower the M-VaR, the less risk the fund is estimated to add to the portfolio. A positive M-VaR indicates the fund is reducing the risk of the overall portfolio when it is added.

Stress Testing
Stress testing is a useful supplement to VaR models, as it allows investors to perform “what if” analyses designed to simulate extreme market events. Using historical changes in global financial markets (equity indices, interest rates, foreign exchange rates, credit spreads and commodity prices), an investor is able to estimate how the market value of a portfolio might change if markets were to rise or fall to the same degree as during a historical event. Popular examples include the ’97 Asian Flu, the ‘94 bond market crash, the ‘98 Russian debt crisis and, of course, the recent market conditions of 2007 and 2008. While no market event ever repeats itself in the exact way, examining your portfolios under a variety of stress scenarios provides a better understanding of a fund’s potential performance under a variety of market conditions.

Exposure Analysis
Arguably the most important (and often overlooked) aspect of risk analysis, is the simple question, do you know where the bets are being made? Exposure analysis is the process by which investors decompose portfolios into simple measures of where they are exposed to the market. This means capturing the full notional exposure or true exposure by factoring in derivatives. Exposures are typically viewed by different areas (e.g., region, country, industry sector, credit rating, market capitalization, etc.). An extension of exposure analysis, risk factor sensitivity analysis quantifies a portfolio’s sensitivity to isolated changes in market risk factors, such as equity indices, FX rates, credit spreads, interest rates and commodity prices.

Leverage
A common theme through much of this past financial crisis has been leverage. One easy way to understand a fund’s leverage is by quantifying its gross exposure—calculated as the sum of the market values of a fund’s long and short positions, including exposure from all derivatives. Whether expressed as an absolute number or a percentage, gross exposure is also useful for monitoring a manager’s compliance with stated risk or investment parameters and assessing style or strategy drift.

Summary
Risk management is all about using the most complete information to make the best investment decisions. Investors should utilize a variety of tools to properly monitor risks and exposures and gauge performance under normal and extreme market conditions. A strong investment process is built around a strong risk process, the foundation of which should be a holdings-based risk analysis.

Measurisk is the leading provider of portfolio construction and risk analytics for hedge fund investors. Measurisk, which had previously been associated with Bear Stearns as Bear Measurisk was acquired during J.P. Morgan’s acquisition of Bear Stearns and renamed Measurisk. It is now part of the Treasury and Securities Services family of businesses serving institutional investors of hedge funds.

For more information, please visit our Website at www.jpmorgan.com/wss