Response to Comments made by ITA member (John Shelton)

Ice Flows – Must be getting close to the Alaska Glaciers

In reaction to my recent Rutherford Review post (https://itawealth.com/rutherford-portfolio-review-tranche-5-21-january-2022/ ) John Shelton commented and asked the following questions:

“Thx for the very gloomy report.

From a psychological perspective, what influence do these trend parameters have on the BHS buy and sell plans? Those of us who study such things know that the Recency Bias would cause all of us to be overly influenced by what has happened last week. The risk is to become overly defensive and maybe consider buying something like SH to short the market. Are you able to stick to the BHS recommends or do they nudge you to alter your judgment in some subtle ways? Can you share with those of us with feeble minds what mental tricks you use to keep cool and stay the course?”

I have chosen to reply to these questions in a separate post for 2 reasons:

• My explanations may be of interest to other ITA readers who might have the same or similar questions;
• I spent a couple of hours writing a response yesterday (in the comment section of the referenced post) and lost it when I was automatically logged out of the site before I could publish it.

I can’t remember exactly what I said yesterday (some sort of cognitive defect) but here’s my responses to the questions today:

“what influence do these trend parameters have on the BHS buy and sell plans?” 

Ideally, none. However, this assumes that the model is perfect – and it is not – no investment model will ever be perfect under all market conditions.

At it’s root, the BHS system is built on a momentum model that uses measurement of both “absolute” (time series) momentum and “relative” (cross-sectional) momentum in much the same way that Antonacci uses these measurements in his simple Dual Momentum (DM) models. For an investor with a long-term time horizon (say, greater than 20 years) these simple systems have worked very well (based on back-testing using historical data).

However, with the recent (over the past 10 years) widespread acceptance of the momentum anomaly and the inclusion of the “momentum factor” in “Smart” Funds (ETFs), there is some evidence that the “momentum effect” may not be as strong as it once was. Momentum is classified as an “anomaly” because it does not fit in with the theory of the Efficient Market Hypothesis (EMH) which assumes that current prices are fair because they reflect the absorption/considered impact of all available information. There are arguments as to why the “anomaly” exists that is generally split between those believing it to be a behavioral phenomenon resulting from cognitive biases (the anti-EMH view) and those classifying it as a risk premium (to, more closely, comply with the EMH). There are also suggestions that it is just a product of data mining. I tend to believe more in the behavioral arguments – but, in the end, I don’t really get too worried about the root cause provided that there is continuing evidence that an “effect” exists.

However, depending on the look-back periods used to measure “momentum”, the simple systems can be slow to respond to changes in market conditions and may or may not work as well in the future. In addition, as we get older, we may not have the luxury of time to recover from large draw-downs. For these reasons I have tried to add shorter-term signals to the basic DM model in an attempt to provide faster reactions to changing market conditions and lower draw-downs. This is not a trivial task since momentum (at least in the DM context) is, essentially, an intermediate term (1-12 month) holding strategy. In the shorter- and longer-term time frames, prices tend to be mean reverting – that is the antithesis of momentum.

The Heikin-Ashi (HA) signals that I am using in the BHS model are being used as short-term momentum/trend signals (that can be visualized in the HA_Charts) on top of the more conventional absolute and relative strength momentum measurements to generate a “Score” on which to rank investment opportunities. This is probably about the best that I can come up with – but I recognize that adding the short-term signals may result in whipsaws due to the possibility of mean reversions. Although, in an ideal world, it would be nice (easy) to have a purely mechanical system that is always right, this is not realistic and I believe that there are times when we need to make discretionary decisions (that may eventually be better or worse than sticking with the mechanical recommendations – depending on the extent to which our cognitive biases may get in the way of our sensible reasoning).

With this in mind I think that it is prudent to (occasionally) take a look at the markets from a different (and maybe wider) perspective. Maybe this helps us make better discretionary decisions?

“Recency Bias would cause all of us to be overly influenced by what has happened last week”

This is very true – so (my “trick”) I always remind myself that there is a (high) probability of mean reversion – therefore maybe I need to wait for confirmation of the (short-term) trend through continuation. This is difficult to build into a model and just adds another parameter (lag) that can be adjusted to “optimize” a backtest so as to give us (false) hopes of success. Therefore, I think the decisions have to become discretionary at this point. If we have positive short-term signals on top of positive long term signals I will always accept the recommendation. This occurs when we see Scores of 9 or 10 in the Kipling workbook. Note here, that a Score of 9 means that we do not have positive absolute momentum from the longer term momentum measurements – I can tell this because of the relative weightings of the signals – and the strict DM rules are violated – but the two positive short-term HA signals override the requirement for longer-term positive momentum. We are essentially “bottom fishing” – so maybe should be considering tighter stop-loss orders (if using them).

“The risk is to become overly defensive and maybe consider buying something like SH to short the market.”

Unless SH (or other short ETFs) were already in my portfolio quiver (previous “sane” choice/decision made while not in panic mode) I would not add it just because HedgeHunter was pointing out that the market was looking a little weak 😊

Personally (but not necessarily correctly) I don’t like including short (inverse) ETFs in my portfolios because:

1. Equities go up over the long term (that is what investors should be focusing on) and we get paid for taking the risk of holding stock (long) – harvesting the risk premium;
2. Daily rebalancing (for ETFs that “track” an index) suffer “drag” from the rebalance requirement/process that adds an additional head-wind.

This is not to say that we could not make (good) money in a bear market when prices are going down in a smooth trend – but this is not the norm (we haven’t really seen this since 2008) – 2020 was a “panic crash” that was difficult for an “investor” to identify in time to make a sensible decision. So, I won’t say that I would never use them (never say never) but it would be very rarely, and probably only in small size.

“Are you able to stick to the BHS recommends or do they nudge you to alter your judgment in some subtle ways?”

As you’ve seen from my reviews I can deviate slightly from the BHS recommendations – although I try to stick with them as much as possible. Rather than being deviations from the BHS model my other (non BHS) adjustments are likely to allocate unused Cash (not called for by the BHS model) to assets suggested from the rotation data – but I haven’t managed to come up with a pure rotation model that outperforms the BHS model (in back-test) so these decisions are discretionary.

I hope this answers the questions asked and that this post clarifies some issues for other readers.

David