The Alpha moderate risk portfolio framework

Moderate risk doesn’t mean what many investors probably think it does. It’s hard to label this type of strategy, which is bolder than running a ‘balanced’ portfolio but more conservative than an ‘adventurous’ one. Asset Risk Consultants (ARC) counts portfolios that fall between those two stools as “Steady Growth”, but we’ve called them ‘moderate risk’.

ARC bases its classifications on the risk of a strategy compared to equity risk: for example, portfolios that are 60-80 per cent as risky as equities are sorted into the Steady Growth bucket. Considering the MSCI World index fell 55 per cent in the global financial crisis, this implies that investors should be psychologically prepared for drops of roughly 30-40 per cent in portfolio value when times are bad.

It doesn’t necessarily follow that such a loss is likely, but peak-to-trough drawdowns are the real-world risk people care about most, so shouldn’t be sugar-coated. Some blends of asset allocation are better than others at mitigating these worst spells. The Alpha moderate risk system suggests a strategic allocation above 60 per cent to be invested in shares, but going back to early 1978 its maximum drawdown is 27 per cent.

This fall occurred in the early-2000s bear market that followed the dot-com bust, a period when the MSCI World index dropped 48 per cent. But investing isn’t about avoiding risk, it’s about managing it, and for those who have the time to see their portfolio recover from any major shocks (ie won’t be forced sellers who must crystallise mark-to-market into actual losses) the moderate risk strategy could be appropriate. After all, over time it offers realistic prospects of a smoothed-out annualised rate of return of 7.5 per cent after inflation.

On its own, the moderate risk SAA is a good way to passively invest, dilute the worst risks and keep stress down to rebalancing to target allocations once or twice a year.

In practice, investors making regular contributions would do better than the risk-adjusted returns of the static asset allocation weights we assume in calculating risk, due to the pound- or value-cost-averaging effect of spreading new cash inputs over different valuation levels. It is possible to do better, however.

The asset weights in our balanced SAA are high level and there are adjustments within asset classes that would make a difference. Changes like these are known as tactical asset allocation (TAA). For example, in 2022, investing the bond allocations in funds with a shorter average duration (which would lower the risk of capital loss from rising interest rates) would have made the portfolio performance much more robust.

Hindsight is a wonderful thing, of course, but often a good strategic asset allocation cushions the early stages of volatile markets and buys investors time to make tactical adjustments to lessen overall downside. There are also times when some of the riskier allocations in the portfolio can be changed to focus on tactical assets such as commodities. In short, the SAA provides a useful management framework.

How we decided on our moderate risk asset allocation

The Alpha ‘Moderate Risk’ strategic asset allocation was developed using indices tracking the total returns of seven core asset classes back to the beginning of 1978. These were UK shares, developed market international shares (hedged to sterling to eliminate currency effects), UK government bonds (gilts), US government treasury bonds (but unhedged to give the US dollar exposure), UK real estate, diversified commodities (with a sterling hedge) and gold (also hedged).

We applied a model that worked out the risk-to-reward ratio of owning each asset. This didn’t just focus on volatility (measured using standard deviation of asset returns from their mean value), it also adjusted for what are known as skewness and kurtosis.

Unlike data that follows a normal distribution, which displays as a symmetrical bell curve on a histogram, asset returns are often negatively ‘skewed’ ie, there are frequent small gains and the losses, although fewer, are larger. Kurtosis refers to the fatness of tails in a frequency distribution, with the high-magnitude losses that occur resulting in fatter tails on the negative (left) side – known as leptokurtic distributions.

Next, we used a coefficient – ‘theta’ – to estimate the level of utility investors with different levels of risk tolerance gained from owning each asset given its risk/reward profile. Investors with the lowest risk tolerance get the least utility from owning an asset with high left-tail risk, even if it might also achieve a high rate of return smoothed out over time. These conservative investors are assigned a greater theta for the utility calculation.

In our four strategic asset allocation models, the theta coefficients in ascending order are: Adventurous (2.5), Moderate Risk (3.5), Balanced (4.5) and Cautious (6).

Finally, we used an optimisation solver to calculate which combination of our seven assets gave the highest utility given the theta we had assigned to different risk groups. In the case of all our strategies, the optimal risk/reward portfolio eschewed diversified commodities and UK real estate as long-run strategic assets.

Below is the long-run strategic asset allocation for a moderate risk portfolio, based on the opportunity set of investments that we surveyed:

How we made the long-run Moderate Risk SAA contemporary

Going back to 1978, it is realistic that there would have been a big home bias for investors in shares, with even the largest overseas stocks more difficult and costly to acquire. The reason UK shares feature so prominently in our asset allocation models is the All-UK market index, although naturally concentrated in large-cap shares, has more exposure to smaller companies than the MSCI World index that we use for international shares. Although in the recent past, big companies especially in the US have made spectacular gains, over the 45-year period surveyed the UK index has benefited more from size and momentum premiums.

While a slight domestic bias makes some sense – especially for income investors wanting to benefit from the UK plc dividend culture and to receive those payments in pounds – the modern equity investor needs to think globally. This is now much easier to do thanks to low-cost exchange traded funds (ETFs), which are the tools we use to create a contemporary and easily investable version of our strategic asset allocation.

We keep true to our risk-management principles in two ways. The first is to not alter the levels of bonds from our long-run SAA model and to set an upper limit for the gold allocation of 5 per cent. This means any changes are made by altering the balance between the riskiest assets (shares and gold within its additional constraint).

Secondly, we calculate the risk budget for the long-run strategic asset allocation and stipulate that the adjusted strategic asset allocation cannot exceed this. The measure used for the risk budget is related to Modified Value at Risk (MVaR), which also takes account of skewness and kurtosis.

Using daily returns for ETFs tracking the five asset classes that made it into our strategic asset allocations (UK shares, global shares, UK gilts, US treasuries and gold), we applied the weights of our long-run balanced asset allocation and calculated that we would expect to lose a minimum of -2.67 per cent in portfolio value on the worst 1 per cent of days (we could also estimate the average extreme loss on those days but for our purposes we are concerned with the risk threshold).

Then, we look to optimise the level of returns we can achieve, subject to the MVaR threshold being no worse than -2.67 per cent and the constraints set around keeping to our bond weightings. We do this by running a query to solve for the highest MVaR Sharpe Ratio.

This ratio (designed by Favre & Galeano) is a variation of the famous Sharpe Ratio. The numerator of the original is the excess return of a security or portfolio over and above the ‘risk-free’ rate of return (the yield on a safe sovereign bond) and the denominator is the standard deviation of the security/portfolio returns.

The MVaR Sharpe Ratio has the same numerator but the denominator is the risk-free rate minus the risk threshold we calculated. Running a query to achieve the highest ratio subject to the outlined constraints, the weights suggested are used for our contemporary strategic asset allocation for a moderate risk portfolio.

For all of these models, we owe a mention of gratitude to NEDL – the channel focused on teaching coding solutions for investors.