Investment managers’ fees need to be justified in two ways: managing the risk in client portfolios and making calls in search of alpha. There is a balance to be struck. Without being reckless, managers should be willing to deviate from benchmark allocations. In a world of robo advisers, where risk-evaluated portfolios can be constructed quickly and cheaply using exchange traded funds, clients rightly question the value of portfolio managers and active funds that simply track their benchmarks.
Handsomely paid managers should be prepared to raise their heads above the parapet and take a position on which asset classes and securities will outperform. There are scientific ways to monitor investment risk and assess how this plays against the potential rewards. For the IC Alpha portfolios, selected as part of our Asset Allocation Overview, we have used SmartRisk software by U.S. firm Covisum. We ran our strategic asset allocations for riskier (Steady Growth) and balanced investing styles and for our tactically adjusted versions of the portfolios.
The SmartRisk system is based on ‘fat-tail’ statistical modelling which recognises asset returns are not normally distributed. In other words, it assigns more realistically high probabilities to larger negative returns than the still widely-used Gaussian model, which assumes asset returns of various magnitudes are symmetrically split about the mean value i.e. in a bell curve. Working out probabilities for parameters which have given rise to past movements in asset prices, SmartRisk makes confident predictions about the average loss an investor might suffer in the worst 0.5 per cent of periods. SmartRisk could be thrown by a never-before seen event (a “Black Swan” in statistical vernacular) but overall, it gives us a more accurate estimate of potential downside for our portfolios and shows whether tactical changes add or reduce risk.
The reports we ran estimate the average size loss our strategic and tactical asset allocations might suffer in the worst 0.5 per cent of months. This is not to say we’d only experience big losses every 200 months (sixteen and two-third years), in the 2008 crisis bad months happened close together. The SmartRisk figure does, however, give a strong idea of the monthly losses to be comfortable with in a very turbulent period, before embarking on a given investing policy.
Risk (average expected loss in the worst 0.5 per cent of months) for our strategic and tactical asset allocations
Portfolio | Monthly risk (SmartRisk) | Monthly risk (Gaussian) | Reward to Risk | Asset Interaction |
Steady Growth SAA | 12% | 7.3% | 1.4 | 21 |
Steady Growth TAA | 11.2% | 7.4% | 1.3 | 35 |
Balanced SAA | 8.1% | 5.3% | 1.6 | 30 |
Balanced TAA | 9.4% | 6% | 1.4 | 38 |
Table Source: Covisum
A glance at the table shows Gaussian models considerably understate the risk of loss and that our portfolios are only suitable for investors who would not be fazed by losses of, on average, 8-12 per cent in the worst 0.5 per cent of months.
Reassuringly, our tactical asset allocation tilts haven’t significantly added to the risk inherent in the strategic benchmark allocations. The riskier steady growth portfolios have a lower reward-to-risk. This is the ratio of average sized positive gain on days the portfolio is up, compared to the average sized loss on days that it is down. This shows us that, going on past performance (which is not a guide to future), balanced asset allocations have been more risk-efficient i.e. it has taken relatively fewer losses to achieve each unit of return.
Should we, therefore, simply abandon the riskier strategy in favour of a more balanced approach? Not necessarily, as the better reward-to-risk for balanced strategies, which have a higher allocation towards bonds, were achieved in a period where bonds outperformed equities. In the current environment of rising interest rates and tapering/unwinding of central banks’ quantitative easing policies, bonds are not going to repeat their spectacular price gains since 2006 (the start for data sourced for the model portfolios).
The most important figures from the risk reports are the headline SmartRisk percentages. The reason to follow a balanced strategy is because a monthly portfolio loss above 10 per cent is unconscionable (it is still possible but more of a genuinely extreme outlier). Reward-to-risk below one is, of course, a warning that too much risk is being taken relative to potential rewards and that the whole investing policy needs to be rethought. For our portfolios, we could infer that although the balanced allocation has been marginally more risk efficient, the difference is not huge; it’s likely that better relative performance for equities versus bonds would see the steady growth portfolio becoming the more risk-efficient in future.
The preference for equities over bonds drove our tactical asset allocation choices. We added risk to the balanced asset allocation by moving from 45 to 50 per cent equity holdings, including a doubling of the emerging markets position. In the case of the steady growth portfolio, the total equity allocation was unchanged but a greater weighting was also given to EM. With both portfolios, the exposure to U.S. stocks is reduced relative to other developed markets. Risk was taken out of the steady growth portfolio, as the 20 per cent bond allocation was split between more shorter duration government bonds, in acknowledgement of the inverse relationship between rising interest rates (and therefore bond yields) and bond prices which poses a threat to the capital value of bond funds.
Our active choices could prove to be wrong. Equity markets have responded nervously to the tough rhetoric on trade being exchanged by the United States and China. Protectionism would be bad for shares and we may need to re-assess tactical tilts when next reviewing the portfolios. We cannot know exactly how disputes on tariffs will play out but, having checked a robust measure of the risk we are taking on, we are much more informed about the magnitude of potential downside. In any case, tilts from our strategic asset allocations are not extreme and the rising rate environment that greatly affected our choices is still the primary consideration.
The last element of the risk report Covisum ran on our strategic and tactical portfolios was the asset interaction. This shows what percentage of the sum idiosyncratic risk of assets is diversified away by holding the portfolios in the weights specified. Both sets of tactical tilts from the strategic portfolios have a higher asset interaction. This, however, does not equate to a lower overall risk for the balanced TAA compared to the balanced SAA.
The reason is SmartRisk attempts to quantify how bad things can get for a portfolio and diversification scores measure how assets inside the portfolio interact together in such a scenario. Whereas we use the MSCI World Index for global equities in the strategic asset allocations, for the tactical tilts we have used a different combination of indices, to change the balance of our equity exposures. The combinations of indices used in the balanced TAA have less positive correlations but the TAA is still riskier in aggregate than the balanced SAA.
With our tactical allocations, we are making the call on diversification enabling us to take advantage of greater dispersion in asset returns in a post-QE era. Again, the most important statistic is the SmartRisk number; we want to assess how much risk we are taking in pursuing our active portfolio choices. For a balanced portfolio, we are slightly adding risk because we believe a different blend of investments will do better over the next few quarters, compared to the past twelve years we have studied. The same thought processes informed the steady growth decisions and it is perhaps unsurprising that the risk profiles of the balanced and steady growth policies have converged somewhat given our bearishness on longer duration bonds.
How would our asset allocations have performed in the past?
An examination of the past performance of strategic benchmarks gives rise to questions about the best ways to invest passively. Our strategic asset allocations are intended as benchmarks for tactical asset allocation decisions. By that we mean risk benchmarks. As accurately as possible, the level of loss we are prepared to stomach is set and the strategic benchmark is a group of investments that reflect the universe of securities we might choose from. For a simple passive strategy, however, you might want a slimmer portfolio.
Had you invested £100 in the MSCI World Total Return Index on 03 January 2006, it would now be worth about £225, in nominal terms the annualised rate of return has been about 6.8 per cent. Take inflation into account however (annualised UK RPI has been around 3 per cent over the period) and the real rate of return for UK investors in the global equity index has been 3.7 per cent in the past dozen years.
This period took in one of the most dramatic equity bear markets in history and at the MSCI’s lowest ebb, in March 2009, the original £100 from 2006 would be worth only £58 in nominal terms. This was a drawdown of £76, or close to 57 per cent, from the pre-crisis peak of £134.
£100 invested in either of our strategic asset allocations would now be worth about £174 – a nominal annualised rate of return of 4.6 per cent, since 2006. Taking RPI as the measure of inflation, this translates to a real rate of about 1.6 per cent. These lower returns, compared to the MSCI, would have been achieved taking less risk. The steady growth £100 would have been worth £76 at its lowest point, a 39 per cent nominal fall from its previous high. The balanced SAA saw a lowest level of £84 – about 28 per cent off its pre-crisis peak – making it the more risk-efficient of our two strategic allocations.
If we were to apply the same test to the tactical asset allocations, then we would have seen paltry annual real rates of return since 2006 of 0.63 per cent for the steady growth TAA portfolio and 0.72 per cent for the balanced TAA. This misses the point of the TAA portfolios, however; they are not selected because of how they would have done in the past, they are the adjustments we hope will outperform over the next 3-12 months against today’s economic back-drop. Therefore, the main concern is potential risk, not back-tested simulations of tilts we would not necessarily have made in past circumstances.
Estimating risk is easier than basing forecasts of returns on past performance. After a decade of ultra-loose monetary policy many investments are expensive. We cannot assume the MSCI World for instance, with its heavy weighting towards the U.S. mega-caps, will continue to outstrip the indices in our TAA portfolios.
With the benefit of hindsight, we could say it would have been better to invest in the MSCI World but this doesn’t account for the psychological distress of seeing over half of all capital eroded in the financial crisis. For some investors, who had a steady income and no need to draw funds and crystallise losses in the bad times, holding an MSCI World tracker was an option if they could hold their nerve.
The reason for active investment is the belief it is possible to achieve higher returns for the same amount of risk. Or take less risk to achieve more modest targets. The point of our Alpha Asset Allocation portfolios is to provide a risk-based framework for blending investments into a cohesive long-term strategy. For investors who don’t believe this is possible through active choices, then the best option is to search for the simplest combination of tracker funds that deliver the best return for risk.
As well as tracking the performance of portfolios based on TAA adjustments, we’ll soon create versions where we have added securities flagged by the Alpha screens to the appropriate slice of the asset allocation. We’ll also factor in the impact of charges and the frequency of rebalancing. These will be genuine active portfolios, not closet trackers of the strategic risk benchmarks.
