When we decide to “invest” there are a few very important decisions we need to make:
- What assets will we choose to include in our portfolio;
- What model/system will we use to manage our portfolio;
- How will we allocate funds to the assets in our portfolio.
In a recent post Lowell offered some ideas for members that might be interested in managing a “Robo” portfolio based on ETFs used by Schwab in their “Robo Advisor” portfolios. This post identified a list of nineteen assets and suggested the use of the LRPC momentum system found in the Kipling workbook. Lowell’s portfolio looked like this:
In this post I will focus on asset allocations. The Kipling workbook allows us to manage allocations in two ways – the first is through the use of the Maximum Asset Allocation (AA) column shown in the third column from the left in the above screenshot. Lowell suggested a range of maximum AA’s from 5% to 35% that can be set based on investor preference. These may be based on allocations normally found in Strategic AA (Buy-And-Hold) portfolios or simply on personal comfort level or preferences.
In Lowell’s example he places no limitation on the number of assets to be held (provided that they satisfy the Buy requirements of the LRPC model) so we have a Tranche sheet that looks like this:
The second place to (indirectly) define allocations is in the Position Sizing sheet:
where Lowell suggests setting the Maximum Portfolio Risk (MPR) to a personal “comfort level” – in his post Lowell suggests a reasonable value of 6% – as shown in the above figure. Since there are 14 Buy recommendations this means that the single Maximum Trade Position (MTP) risk calculates to a low 0.43% risk in the above figure. At this level we see that two assets would exceed our risk level unless the investment allocation is reduced as seen in the Allocation To Cash column. Note that in the above example the Suggested Portfolio Risk (SPR) calculates to 4.84% – well below the target 6% level.
So let’s see how Lowell’s portfolio performs over the period from 29 February 2016 – the earliest date I can start based on available data for all assets:
where we see good performance through 2017 but sideways performance over the past 2 years. 2018 and 2019 have been tough years for momentum systems generally (not just the ones available within the Kipling workbook). However, I don’t think that too many of us would be too disappointed with a ~30% return over the past three and three quarter years.
However, the focus of this post is on asset allocation rather than performance per se – so let’s take a look at some alternatives.
My first change might be to change the Portfolio risk calculations in the Position Sizing sheet. My (personal) preference is to set the MTP risk to 2% (the default value when the workbook is downloaded). When we do this for the Robo portfolio we see this:
Where we see a scary number for the Maximum Portfolio Risk – 14 x 2% = 28%. But this is not a realistic value – Suggested Portfolio Risk is only 5.01% – slightly above the original 4.85% but still well below our desired/preferred 6% level – and we are fully invested (no Cash position). So my preference is to only change the MTP risk (to less than 2%) if my Suggested Portfolio Risk rises to an unacceptable level. This, generally, keeps me fully invested.
So, let’s take a look at how this changes performance:
and we see (yellow line) a consistent improvement – with an increase in total (4-year) returns of ~4%.
As noted above, risk is also affected by the Max AA values set in the portfolio sheet. Obviously there are millions of possible combinations – and one combination will probably generate the best performance/returns. However, we have no way of knowing what this combination might be so (again my personal preference) I will set all Max AA’s to 100% (i.e. no restrictions). This means that the total available funds could be 100% invested in a single ETF – and many investors may be uncomfortable with this. However, remember that excessive risk can still be managed through the position sizing sheet. So, let’s take a look at the portfolio with no Max AA limitations:
and we’ll keep the 2% MTP risk set at 2% as used above. Running a back-test using these parameters we get the following picture:
where we see (green line) a significant improvement in total returns to ~40% over the 4-year period.
Although the maximum Suggested Portfolio Risk (SPR) tends to lie between 4% and 8% – the average is still probably less than 6% and can be throttled back if necessary.
Although not directly related to asset allocation, another decision we make is the maximum number of assets to be included in the portfolio. In all the above tests/examples we placed no limit on Maximum Funds (MF) – with this set to 20. However, let’s take a quick look at the impact of reducing MF to 7 (~1/3 total assets available):
Over this 4-year period we see that reducing the number of assets (purple line) is beneficial in strong up-trending markets (2017) but does not do as well in consolidating sideways markets and exhibits more volatility.
Finally, although I’m sure that most of us would have been happy with 30-40% returns over the 2016-2019 period, let’s put this in perspective with a 4-year Buy-And-Hold of the S&P 500 (SPY):
Pretty frustrating for momentum investors and shows that Buy-And-Hold would have been the best strategy over this period – but, remember, this provides no protection in strong downturns.
There is no perfect answer to the question of “What is the “best” asset allocation model?” and, over the years, many options, including momentum ranking, Minimum Variance Optimization, Risk Parity and a number of complex combinations have been described on this site and included in various versions of the Kipling workbook. However, differences are difficult to quantify over periods of different market conditions – so I have always said “If in doubt, equally weight”. The current workbook does essentially this, although use of the position sizing sheet does build in an element of Risk Parity. This post is intended to provide a little more insight into the tools available within the Kipling workbook for risk management through asset allocation – so I hope it helps.
Update: 22 May, 2020
In response to a request from Jack B, I have updated the plots generated for the above post. These EOM plots now cover the 2020 market decline resulting from the Covid-19 pandemic:
As can be seen, none of the allocation strategies fared well through this period – and reinforces the significance of timing (Review-Date) luck. At the time of the February review the decline was only ~10% and was not sufficient to register a Sell recommendation for the LRPC system. In a post near the end of February (https://itawealth.com/is-the-use-of-stop-loss-orders-a-good-strategy-tactic/) I reported that my stop-loss orders were getting hit ~23 February – and this was a good time to get out of the market, but difficult to detect without close attention to the market and still difficult using look-back periods designed for longer term “investing”. By the end of March we had seen a 35% decline in US markets.
I mention the significance of timing luck – check out the recommendations for the Rutherford Portfolio on 6 March (https://itawealth.com/rutherford-portfolio-review-tranche-2-6-march-2020/) where Sell signals were generated.
As can be seen above, the strategies using fewer assets (purple and dashed purple lines) managed to perform slightly better – but still not well. Charge one up to the benefit of stop-loss orders or more frequent reviews on this one.