You can have anything, but not everything, you want

One of the more challenging aspects of helping clients solve their investment related problems is figuring out their true investment objectives. Below is a made-up conversation with a fictitious client that exemplifies the challenges presented when working through an investors’ objectives:

Q: What would you say are your key investment objectives?
A: We want to generate a return of 5% per annum (over 5-year periods).
Q: Ok great, is that before or after inflation?
A: Not sure. Does it make a big difference given inflation is so low? I suppose we do want to preserve the real purchasing power over time; so, after inflation.
Q: And how would you describe your risk tolerance?
A: Fairly high; the investment committee understands that returns only come with taking risk so we are comfortable.
Q: Would a drawdown of say 20% in a given year be too high?
A: YES! That would be unacceptable. Ideally we would like to see positive returns over 12-month periods. However, we could probably tolerate a 10% loss in a given year.
Q: Understood. Do you have an income target?
A: 4% income per year would be ideal as we expect many of the underlying investors to be approaching retirement.
Q: Are there any other expected liabilities or cash flow requirements?
A: Possibly. Occasionally we have projects that need to be funded in the order of 5-10% of the total asset size.

The most notable thing about this mock Q&A is the sheer number of objectives uncovered over the course of a very short conversation. When working with clients, existing or prospective, we almost always determine one or more ’hidden objectives’ through our discussions and analysis. Furthermore, we often determine that two or more of the objectives can be contradictory. For example, an aggressive return target of 10% while keeping volatility below 5% is clearly inconsistent. Often a client’s extreme aversion to a real (after-inflation) capital loss would warrant allocating 100% of a portfolio to inflation-linked bonds, yet their return targets remain relatively ambitious and out of reach.. This solves one problem but creates others.

The bottom line is that we can likely build a portfolio and strategy to successfully achieve any one of these objectives but find it challenging, if not impossible, to achieve all objectives simultaneously. Investing is like life in this respect as the self-help gurus often preach: you can do ANYTHING but not EVERYTHING.

A VaRy good idea?

Good risk management typically starts with good risk measurement. As 20th century management consultant and business influencer, Peter Drucker, famously said, “You can’t manage what you don’t measure.” There is now a relatively common set of metrics most portfolio and risk managers use to assess the ‘riskiness’ of a portfolio. The inverted commas intentionally reflect the inherent limitations with such risk measures, which are often based on the flawed assumption of normal distributions and rely on historical data. Statistics such as standard deviation, value-at-risk, and drawdown are useful but only form one piece of the overall risk measurement mosaic. Other techniques such as stress testing a portfolio or determining how a portfolio would have behaved during a particular market scenario have become the new ‘table stakes’ within risk management.

This then begs the question: what risk management problem are we trying to solve? We can answer this by more clearly defining ‘risk’ which in the context of a broad investment portfolio is failing to meet the investment objective(s). And since we’ve already established that most investors have more than one objective, we need a way to maximise the chance of delivering on all of these multiple often-contradictory objectives. To help solve this problem, we have developed a systematic method of maximising our chances of success (or minimising the probability of failure).

And despite many people’s worst fears, mathematics is essential in elegantly solving this problem.

…Introducing the Weighted Risk Metric

In order to be able to make an informed decision and address the trade-offs inherent in achieving most investors’ objectives and risks, we invented our proprietary Weighted Risk Metric, which incorporates the specific requirements in the calculation of an overall risk score for each candidate strategy.

The Weighted Risk Metric is a mathematical concept developed by our Multi-Asset Solutions team. It calculates a Weighted Risk Metric score for each portfolio on the efficient frontier which is a measure for not achieving the combined objectives. The Weighted Risk Measure can combine (or weigh in) different risk measures such as shortfall risks or conditional value-at-risk (CVaR) for combinations of target returns and investment horizons. The model determines the risks of not achieving the specified ‘aggregate’ goal. When the scores for all portfolio are known we then simply select the portfolio with the lowest score which has the best chance of successfully delivering on the combined objectives.

Revenge of the nerds - Balancing risk/return to solve all your problems

In a traditional strategic asset allocation (SAA) approach, a portfolio is constructed by optimising the asset universe available within a portfolio based on forecasted risk, returns, and correlations. As many balanced portfolios implement a standard SAA approach and use long run or equilibrium capital market expectations, we often see the resulting asset allocations look remarkably similar, as they only attempt to optimise one return or risk objective over a single horizon. In reality, investors do not have an unlimited time horizon and will typically only invest over a specified investment period.

The example below provides the output generated from running a traditional asset allocation analysis. In this instance we are targeting a total return of 5.9% over an investment horizon of 5 years, or, a 4% return in excess of inflation, which at the time of writing was hovering around 1.9%. Portfolio 1, illustrated below, has a 95% allocation to ‘growth’ assets, or equities and a 5% allocation to ‘defensive’ assets, or fixed income and cash.

Source: Colonial First State Global Asset Management

The ‘growth’ weighting is higher than even a typical 60/40 balanced portfolio, due in part to the current low expected return environment as central banks continue to attempt exiting the regime of low interest rates and quantitative easing just as the global economy begins to slow.

While the traditional asset allocation approach provides a good starting point, it ignores other factors that one has to consider when making decisions to allocate capital in the real world. While we deliberately kept the introductory mock Q&A brief, it is worth highlighting that there are many other common considerations one faces that are not normally found on the label of an investment product. These include, but are not limited to: 1-year short fall risk (drawdown), inflation protection and an investor’s own individual time horizon.

To help solve for the additional factors, a multi-dimensional framework will provide greater insights, which is where the Weighted Risk Metric comes in. Portfolio 2 below has the same objectives as Portfolio 1, however, we are now incorporating some of the hidden objectives that we uncovered earlier, and assigning weights to capture short-fall risk over any 1 year and inflation protection over 3 years periods. We have still assigned a weight to achieving a total return target of 5.9% over the full investment horizon of 5 years but are no longer solely solving for this objective.

Adding these additional objectives significantly changes the asset allocation, which intuitively makes more sense as we are now solving for multiple objectives. This is a more conservative portfolio but is also more representative of what the investor is actually expecting in terms of consistent outcomes.

Source: Colonial First State Global Asset Management

This portfolio has a 72.5% allocation to ‘growth’ assets, or equities and a 27.5% allocation to ‘defensive’ assets, or fixed income and cash. The portfolio is also more diversified across asset classes than the previous example.

While Portfolio’s 1 and 2 both are based on the same fundamental view of the economic climate and expectations, they have very different risk and return characteristics due to the additional objectives that have been weighed-in for Portfolio 2.

I’m a model, you know what I mean

Risk management should be fully embedded in all aspects of an investor’s thinking and process, not just an afterthought. Indeed, although sophisticated risk monitoring is critical, we should not lose sight on what risk is: failing to meet portfolio objectives. We are the first to acknowledge that the tools we have developed including the Weight Risk Metric, like most models, rely on qualitative inputs and subjective decision making. In this model, we ultimately need to determine how important each objective is up front and no level of complex maths will help us here.

However, as the British statistician, George E.P. Box said, “All models are wrong, but some are useful.” The Weighted Risk Metric provides a systematic approach to buttress traditional asset allocation methods to capture the subtle nuances within investment objectives. Maybe we can get MORE of what we want after all.

Want more on asset allocation?