Forecasting is a necessary but challenging exercise in investing. The more complex the financial markets, the more complicated a forecast. In addition, the natural difficulty of forecasting is exacerbated further by typical human responses to complex situations. Scott Wolle (pictured), Chief Investment Officer of Invesco, believes investors must develop a process that avoids the worst of the obvious errors, has hedges in place, and is constantly reviewed for sufficient robustness…
Invesco’s Global Asset Allocation team has analysed the pitfalls of forecasting in times of uncertainty confirming forecasters’ poor record at forecasting complex, long-term events and highly uncertain variables such as economic growth. In a complex, interconnected system such as today’s financial markets, major shifts can build almost imperceptibly and then accelerate suddenly. Unexpected events have a nasty habit of interfering with the smooth operation of the stylised world that we envision. Over the past 10 years, these have included a major terrorist attack, two wars, a financial crisis and a sovereign debt crisis.
Prudence requires us to understand the snares of forecasts and to avoid them through good design. I see three main properties of well-designed systems: 1) structural resilience; 2) frequent, controlled allocation shifts, and 3) continuous review.
Structural resilience refers to some sort of protection against the negative effects of extreme events that are easy to overlook but occur with higher frequency than standard models would predict – and can have a severe impact. Structurally resilient portfolios are characterised by adequate diversification among asset classes that respond differently to different market environments – for example, long-duration treasury bonds can add a measure of protection to equity portfolios during recessionary periods. Diversifying assets must have a high enough weight in the portfolio to have a meaningful impact when their presence is most needed. This can be estimated by simple scenario analysis or through more stringent quantitative methods.”
In a second step, the risk-return profile of structurally resilient portfolios should be optimised through frequent, controlled allocation shifts. I endorse a disciplined approach that makes short-term forecasts based on a framework that incorporates a wide variety of inputs with a high sensitivity to market-based signals. However, he warns that investors should not shift too far from the structurally resilient core portfolio or fully eliminate the hedges in their portfolio.
The final, and perhaps most important, element is the need for an ongoing search for process vulnerabilities and opportunities. The investment process must be continually challenged, with the team weighing the potential benefits and risks of enhancements with reduced exposure to a variety of known behavioural biases.
There is no way to manage a portfolio without any projections. Even those who deny the presence of forecasts within their process have, by definition, some implicit expectations for how the world does and will work – in short, a forecast. Therefore, it is important to know the limits of forecasting – and to navigate around the most dangerous types of forecasting errors.
