The existing Global Equity Factor and Global Bond Portfolios allow asset allocation at the subgroup level between different ‘styles’ and also with cash, all based on a risk rather than a return basis.These can, and we believe should, be used to provide the ‘core’ of a balanced portfolio. The underlying investor can buy and hold, satisfied that the allocation between different ‘styles’ can be made in a timely fashion, as can any overall risk reduction for market risk purposes.
The allocation for a more active satellite portfolio is something that can be covered we believe by our new Multi-Theme Portfolio. Here we take the same approach of allocating between underlying components based on our risk matrices, but instead of using factors, we use themes, specifically a subset of the Thematic ETFs offered by I-Shares. We focus on the five technology related themes, plus an allocation to Emerging Market Consumers and Gold. The risk matrix for the last two years is shown below.
Table 1. Global Thematic Risk Scores Matrix
The idea is that, in exactly the same way as we do for the Global Factor and Global Bond Portfolios (see here for full explanation), we apply our propitiatory risk scoring system to each of the Thematic ETFs as listed in the table: 1) Automation and robotics, 2) digitalisation, 3) digital security, 4) EM Consumer Growth, 5) Global Clean Energy , 6) Healthcare Innovation and 7) Gold. The lower the score, the lower our view of the risk and the higher our asset allocation. If all scores are 1, then we are equally weighted. At the other extreme, if all are scored 5, then we are all in cash. In between those extremes we balance our portfolios according to our measures of risk. Turnover and rebalancing is relatively low – this is an investment portfolio not a trading one – but frequent enough to manage the changing market risk dynamics. We also have a maximum allocation to any individual ETF.
The table shows month end snapshots, but the risk matrix is calculated daily and the model portfolio performance is calculated at close on t+1 after the signal is given.
The first chart then illustrates the performance of the model portfolio over the last year. Note that we are applying the same allocation principles embedded in our other Model Portfolios rather than simply ‘fitting’ a back test. As such we regard it as a form of ‘out of sample’ test. Not surprisingly perhaps the performance benefits from the ability to reduce exposures as risk started to spike in February and March, but also for the conviction to start to re-invest as risks decreased.
Chart 1: Recent Performance of Global Thematic Model Portfolio
Here from the monthly numbers we can see that the drop in February and March were less than the market due to the risk reduction techniques and were also more than offset by the recoveries in April, May and June, highlighting the advantage of the disciplined approach of selling early, but not being too fearful when risk recedes.
On a calendar year basis, the strong growth nature of the portfolio is best illustrated in 2017 when risk scores were low across the board – as illustrated from the longer term risk heat map (below) – and the portfolio was duly invested. Note the low exposure to gold, not only then, but for much of the period). When risk is low, exposure to growth is high. Equally, the high risk period in 2015 and 2018 saw the portfolio offer downside protection relative to the benchmark. Note how EM Consumer growth, while an attractive long term theme is probably the lowest average exposure over the last five years, highlighting the discipline whereby we can ‘like’ a theme, but our exposure will be tempered by its risk.
Looking at the annualised data (note ytd and 3mth will be distorted as a result) we can see that the Model Portfolio, like the other ones we run currently, has return above benchmark and volatility below benchmark, which is obviously what we are trying to achieve.
The longer term chart illustrates the compounding advantages of losing less when markets go down as well as maintaining upside exposure when they do well.
Chart 2: Longer term Performance of Global Thematic Model Portfolio
The final chart illustrates the shift in the ‘blend’ between thematic exposure over time, which is in our view one of the key benefits of this disciplined approach.
Chart 3: Balance of Thematic Exposure over time
We will be adding the Thematic Portfolio to our regular monthly updates of our Model Portfolios.
