Trusteeship is all about risk management and the basic ideas behind any risk management principles are deceptively simple and logical. You define the risks you want to manage and who’s responsible for them. You take action to limit them. You measure and monitor them and take further action when necessary.
The structure is the same whether you are looking at Risk Standards for Institutional Investment Managers for the Risk Standards Working Group (see box), the Group of Thirty’s standards for the use of derivatives or the Myners Review’s principles. Yet, when the four-letter word risk slipped into the list of things that the UK’s Pensions Act 1995 required trustees to consider when documenting their statements of investment principles, there was a certain sense of unease. Actually pinning down a group of trustees’ tolerance to risk is no mean feat.
How can consultants and fund managers help facilitate that very first step in the risk management process – understanding and defining the risks to be managed? The main problem is that most investment risks – the risk of underperforming a benchmark, for example – are influenced by a variety of complex interlinked causes. Getting to the root of exactly what went wrong when something has already gone wrong is hard enough. Anticipating what might go wrong before it does go wrong is even more difficult. Yet this is what risk is all about.
Let’s look in more detail at the risk of underperformance. Behind the fifth principle according to Myners, explicit written mandates should cover all the reasons why the manager has been hired – and might be fired. Again, in theory, this sounds simple. When you pick a manager you write down the reasons why, make sure that they know what is important, and hold them to it.
When we talk to managers about what is important in their selection there is clear consensus that the right answer is their process. But what do we mean by process? What do we write down in the “reason why we hired them” box? How do we tell when it is going wrong and the risks of underperformance need managing?
To help answer these questions our firm has developed SimIAn, a sophisticated technique to help explain why fund managers have performed the way that they have. Using SimIAn we can:
q identify the parts of the process that contribute most to risk and return;
q measure evidence of skill in the investment decisions.
The SimIAn Process Attribution technique involves:
q constructing a detailed definition of the portfolio construction process and breaking it down into its constituent steps;
q randomly generating a large simulated peer group of alternative portfolios, all of which could have been produced by the same process;
q generating further simulated peer groups using variations of that process with one or more steps missed out;
q comparing the results of the simulations with: each other, and the actual portfolio.
In this way we can make observations about the value added and the risk controlled by each step.
To analyse fund managers’ processes in a consistent way, we have developed what we are calling the SimIAn Process Framework. By mapping a fund manager’s process to this framework, we can then analyse each step in the process, to identify where skill is being shown and risks controlled. For example, we identify what we call influences and controls in the process.
Influences are any themes or styles that influence stock selection and portfolio weightings in a positive direction – “choose these stocks” or “be overweight here”. These would be expected to add value over time, but also to give the product significantly different performance compared with its benchmark.
Controls constrain stock selection and portfolio weightings – “don’t forget this” or “don’t be too far away from this”. These would be expected to constrain the portfolios in a way that reduces the risk of underperformance or of volatility against the benchmark
The SimIAn Process Framework diagram shown in Figure 1 illustrates the simulations that we perform. Each coloured box represents a variation of the manager’s actual process. By comparing the simulations we can attribute the performance to the different steps in the process, answering questions such as whether the influences are adding value, the controls cutting down risk and the managers showing skill in stock selection.
Examples of the things that we look for when comparing two simulations are:
q Are the two variations on the process significantly different? This would suggest that a particular step has a significant effect on the performance of the product.
q Does the second process show better average returns than the first? Is it a consistent extra return or is it volatile? Consistent added value is evidence of skill, rather than just luck.
q Does the simulation show a narrower range of returns? Are the returns closer to or further from the benchmark? This can demonstrate the power of a risk control in narrowing the range of returns.
q How close is the actual portfolio return to the simulated results? Consistent outperformance of the simulated peer group may be evidence of skill.
An example chart is shown in Figure 2. The purple shaded area shows the range of most of the returns each month from a process with similar portfolios but ignoring the effect of a particular influence in the process. The red boxes show how much difference adding the influence makes. It also shows how the particular portfolio chosen by the manager (yellow dot) adds value relative to the process as a whole.
The SimIAn technique offers a way that fund managers and trustees can objectively measure how their skills are being used in their process to give the bottom-line performance. As such, it can help trustees write down the skills they are expecting. As a monitoring tool, it can also give them the evidence they need to show that the process is working smoothly.
Equally it provides objective early warning signals when the process is not behaving according to plan. When the risk of underperformance is increased, action can be taken. In line with what Myners probably had in mind, it allows a more sophisticated approach – not just sacking the manager but instead changing the mandate and adapting the process. In this way the manager’s process can evolve to the benefit of the scheme, and the manager
Sally Bridgeland and Matt Livesey are with the investment consulting division of Bacon and Woodrow in London
Risk standards relevant to portfolio construction processes
q Risk management Identify and understand key risks, including the assumptions and vulnerabilities built into a process Set risk limits at portfolio, asset class and individual holding levels.
q Risk measurement Measure and attribute risk and return to quantify the key drivers. Test how portfolios would behave under various conditions including unusual events and changes in key risk factors.
q Oversight Perform frequent independent reviews of managers’ risk policies and controls. Require managers to show that their activities are consistent with their strategy.
Based on ‘Risk Standards for Institutional Investment Managers and Institutional Investors’ by the Risk Standards Working Group