Why investors need a multi-layered approach to risk
The experiences of the 21st century have sent investors in different directions. Low equity returns since the turn of the millennium have prompted a search for new sources of risk, yet the traumatic experience of the 2008-2009 financial crash have also exposed the limitations of traditional measures. This has prompted a search for new approaches. Armed with a new and much broader set of risk measures, we argue that investors should use a multi-layered approach as they exercise their judgement in building portfolios.
Tracking error has long been a cornerstone of measures to determine risk. Like beta, it can be seen as a temperature gauge. Used in this way, such measures can be compared for portfolios with similar benchmarks and can signal material changes in risk. However, tracking error after the fact (“ex post”) will differ from (“ex ante”) predictions as a result of market events and portfolio changes. The problem is that tracking error may not brace investors sufficiently for the range of possible outcomes they are likely to experience.
As an estimate of sensitivity to the market or benchmark index, beta also has its limitations. It may be based on a weak link with past performance, while there are many non-market influences on share prices. Moreover, some of the biggest risks for equities are events that have never happened, or not in recent times, such as a permanent shift upwards in the yield curve, deteriorating US-China relations and the disintegration of the EU.
Standard measures of portfolio risk are ultimately based on volatility forecasts, which are often based on past volatility that can deceive. And the differing time periods used can be a distorting factor, meaning that risk is underestimated over long periods. In any case, if an investment is well chosen, volatility should be irrelevant or even a buy signal for long-term investors.
Given these and other limitations, such traditional risk measures can only ever be a rough guide in “average” circumstances, yet what sets managers apart is largely their performance in unusual markets. Stress-testing and scenario analysis can help to address these issues by showing how a portfolio might react to significant changes in key variables, such as equity indices, interest rates, currencies or commodity prices.
In similar fashion, factor sensitivity analysis can show how a portfolio might be affected by a change in an extraneous factor, like the oil price or the shape of the yield curve (see chart). For instance, our factor analysis shows that, should the yield curve for German government bonds steepen by one standard deviation (based on five years’ data), a given portfolio is likely to underperform its benchmark by around 0.9%.
Scenario analysis involves thinking about the consequences of different sets of occurrences and, sometimes, what might be done about them. It would have been useful in October 2016, for instance, to think through the implications of an “unlikely” Trump victory for markets.
Another recent development is the use of active share, a measure of the extent of the “bets” an active investor is taking against the benchmark. This may make sense, although it can be misleading if used to compare different strategies: an active share of 80% would be high for a pan-European large cap fund but low for a small cap fund.
Realising the limitations of volatility, some fund managers ascribe “risk scores” to individual stocks. These rely on “fundamental” risk factors, such as leverage, exposure to adverse environmental, social and governmental issues and country risk. Even newer ways to look at risk have emanated from the increased availability of computing power at ever-diminishing cost, while the growth of e-commerce and the internet has expanded the array of data from which we can derive insights.
In sum, we believe good risk management is multi-layered. First and foremost, managers need to consider risk on a continuous basis. Secondly, asset owners need to ensure that their managers are taking the right risks by regularly monitoring a variety of measures. Thirdly, asset owners need to be sure that a comprehensive risk management framework is in place. Given the constant evolution of risk management, it is vital that all the necessary layers are in place and that a manager has access to the whole toolkit.
Read the full report
6 pages | 169 kb