A Portfolio of Diversified Assets

This post was born as a quick reaction to the 2022 Golden Butterfly portfolio performance and then morphed into a collection of thoughts about portfolios, diversifiers, cash vs trend, a levered version of the Permanent Portfolio and…probably other things that will hit my brain while I perform the week-long writing process.

The guy behind PortfolioCharts updated all the portfolios he tracks with 2022 performances. Since I discovered his Golden Butterfly strategy, I was intrigued and followed it closely (I wrote a post about it here in 2021). One of the possible reasons it performed so well in the past is its high allocation to gold, 20%; I thought that few investors would try it and even fewer would be able to stick to such an unconventional allocation when gold was heavily underperforming. A kind of “if you thought that being disciplined with value investing was hard, try this”. So, the Bravehearts that managed to hold reaped great profits…well, great risk-adjusted profits. Then 2022 happened:

The real (as opposed to nominal) loss last year was DOUBLE that the previous historical drawdown. The dataming bat-signal was shining bright above the London sky. Is it possible that there was nothing special in the Golden Butterfly, other than a good fit of past data? In theory, the high gold allocation protects the portfolio exactly from the high inflation we experienced; yet, it did not work as well as it did in the ’70s.

This is a great example of why achieving above risk-free returns is easy when you play with the past but hard when you deal with an uncertain future. Anyone can hold a 100% stock portfolio through the 2008 lows…today, because they know how the story ended. As PortfolioCharts correctly evidence, the 2022 GB loss is hardly noticeable if put in the context of 50-year history; it might represent THE reason why the portfolio will continue to produce extra risk-adjusted returns in the future. Even compared to all the other portfolios in PC database, 2022 wasn’t that bad for the Golden Butterfly.

Here is a quick visual comparison of the Golden Butterly vs the standard 60/40, to remind you how great the strategy is:

Or 2022 can be the start of the “out-of-sample” disaster that so many other data mined strategies experienced in the past. Both points are valid, I think the only imperative here is to ask yourself those questions.

To be fair, the Golden Butterfly looks overly impressive from a USD investor point of view. When replicated by people that have liabilities in EUR or GBP, past results are less outstanding [here I am talking about the standard portfolios in PC; there are many ways to slice that cat, from the asset choices to what to hedge from a currency risk point of view]. For non-cowboys, 2022 is not an outlier at all: unfortunately, not because the 2022 return improved but because you find other years with real losses of around 20%.

Defensive Assets

The other day I was listening to the InvestLikeTheBest episode with Tim Urban. [I will not tell you that you should listen to it because I assume you already know]. Tim highlights the importance and value of ‘idea labs’ or ‘genies’, places where dissent is positive and constructive. I was a pretty dogmatic person, either you are with me or against me, but writing this blog helped me to accept (and ideally even leverage) opinions that challenge my priors.

In this spirit, I will NOT reprehend a post from Monevator where they discuss defensive assets alternative to bonds. At least, not as I would have done in the past, eheheh. Monevator provides a lot of data, graphs and their usual, great writing style. Still, there are two aspects that always trigger me:

• I understand their audience is mainly English investors and those (like many others) have a strong home-country bias. You write for your audience. But why, and this is a remark I can do about PortfolioCharts as well, do not start from a GLOBAL portfolio point of view and then tackle the FX risk implications? Why your base case has to be only UK equities and only UK bonds and…only fish and chips for lunch? US Treasuries might lose their ‘safe heaven asset’ in the future, but I would rather model that bucket this way rather than take UK Gilts because “oh, they are bonds and they do not have currency risk” while they lost any resemblances to a save heaven asset a century ago. Wouldn’t a mix of Developed Markets Government bonds be a better proxy AND a better investment? Sure, FX volatility nullifies the low-vol characteristic of bonds, but in the medium (?) to long term it is mainly good volatility, i.e. it protects if you happen to live in a former empire that is now less relevant than all its former colonies. [I will soon have the opposite problem moving to Switzerland. Yet I would prefer to lose on the FX side while investing in a global portfolio, bonds included, than holding 100% bonds of an economy that represent 0.4% of world GDP].
• If I will ever refer to cash as an asset class please shoot me. Feels to me a bit like suggesting that if you want to improve your life expectancy you should not drive. Data-wise it is correct but…it is really what we want to achieve? Kind of, the wrong low-hanging fruit. Yes, it is there…but we can do better, innit? [This has nothing to do with the ’emergency bucket’ so dear to any personal finance guru; that’s a form of insurance, not a Strategic Asset Allocation element. In fact, it is measured in terms of expenses, not assets].

Trend

Ok, in reality my only diss with Monevator is that they did not even mention trend following as a possible alternative. That’s the real reason why I wrote the previous section.

I recently wrote a guest post for a popular Italian personal finance website. Given the very broad audience, I tried to ‘dumb it down’ as much as I could. When I submitted it, the site owner replied with an emphatic “nope, dumb it down more”. Despite our concerted best effort, we still received feedback on the tone of “why the yield of a bond is negatively correlated to its price?”. I safely assume that the same person who writes a comment to a blog has access to Google but…I guess you never know. The issue is not if ChartGPT will become sufficiently intelligent but if there is a way that would compel everyone to use it.

I understand why Monevator would not want to even stare down the trend-following rabbit hole. This morning I listened to the latest episode of Pirates of Finance and my head was exploding trying to follow all the possible variants of trend following (there is so much straddle info you can manage while grocery shopping at Lidl). But then I remembered the Russillo-life-advice “why do you even coach?”. If you run a website that is supposed to educate your audience, should you not at least mention that this asset class exists? It is a complex matter; given their success, they clearly know what their audience wants better than me.

Anyway, if I did not lose you yet, RCMAlternatives put together a great trend-following guide piggybacking on a Meb Faber thread on Twitter. They cover questions like:

• fees and fees transparency
• definition of trend-following, managed futures and CTAs
• volatility of the strategy
• products

Sure, there are not a lot of products available for European investors. But I learned on Monevator about the existence of BHMG, a closed-end fund that feeds into a Brevan Howard macro hedge fund (yes, macro is not exactly equal to trend-following as a strategy but it is a better proxy than nothing). The guys behind DBMF are launching a UCITS version of the fund this year (UCITS = European retail investors can invest, last time I save you a Google search ;)). If inflation continues to stay high and trend continues to be a good hedge against it, I would bet more products will arrive. Or you might move to Switzerland and get access to the US ETFs 😉 Or you might become a sort-of accredited investor under the UCITS rules (even if you would still have many tax headaches in the UK). Anyway, better to know than not.

While many trend ETFs are backed by white papers that show the benefit of the strategy going back decades, their live performance is short. A suitable example is the Trinity (TRTY) ETF from Meb Faber, a fund that targets an allocation to 25% global stocks, 25% global bonds, 35% trend and 15% other asset classes like commodities, real estate, currencies and alternative investments. In below picture, Port1 is the Golden Butterfly, Port2 is the 60/40 and Port3 is TRTY:

It is hard to push the merit of the asset when you offer this evidence. But I also understand that you need at least a full cycle (or as Meb says 20 years) to draw a conclusion.

ASPY vs XYLD

Jason Zweig recently wrote about investors renewed interest in covered call strategies, given the good results they had in 2022. I chronicled my experience with XYLD, the Global X Covered Call ETF, more than a year ago. I am not sure why, but it prompted in my head the following question: how XYLD performance would look next to ASPY, the trend ETF I described a few weeks ago?

Here is the comparison since ASPY was listed:

And here is a longer test using the ASPY index, the backtested strategy behind the ASPY ETF:

Despite employing two very different plans to control risk, in both horizons they end up almost in the same place (again, ASPY the ETF should trail XYLD by c0.8% per year in the 10-year test). But if you look closely, ASPY would have saved you a lot of drawdowns (see 2016, 2018, 2020 and 2022).

I could hardly think of a better picture to describe the power of trend.

A new, leveraged, Permanent Portfolio

The other day @nomadicsammy, the blogger behind PicturePerfectportfolios.com, suggested this elegant and simple two ETFs portfolio:

• 50% GDE
• 50% RSBT

Both ETFs follow the same “return stacking” philosophy behind the infamous NTSX: they employ futures to get cheap leverage and capital-efficient exposure to some assets. Their combination gives the investor the following portfolio:

• 45% stocks
• 50% bonds
• 45% gold
• 50% managed futures
• -90% cash (leverage)

The original Permanent Portfolio was designed this way: four equal 25% allocations to stocks, bonds, gold, and cash. Here we substitute cash for managed futures and levered it up to almost 2x. The switch makes sense because managed futures should (we are still talking about a strategy that is linked to the specific manager(s) employed) provide the same hedge in turbulent times as cash with higher expected returns.

The ETFs launched not a long time ago but I can mimic the past this way against the S&P500 (Port2):

Did someone say stock-like returns with bond-like risk? 😉

This is just a fun exercise but this combo definitely has potential [the managed future allocation to PQTIX might not reflect the strategy employed inside RSBT and my cost of leverage proxy might not be correct].

How do you plan for the worst?

PortfolioCharts uses baseline returns (the 15th percentile inflation-adjusted compound annual growth rate (CAGR) for a given investing duration looking at every start date we have access to) instead of past arithmetic returns to better assess what a portfolio might return in the future. It is a sensible approach because it gives more relevance to portfolios that exhibit less volatility (and fewer outliers) in their past returns. Similarly, Kris Abdelmessih explains in “Well what did you expect?” why CAGR is a better predictor for a portfolio’s future returns than the arithmetic mean (again, it is a matter of volatility).

High returns, especially if compounded for a long time, are the best ingredient to grow your savings and meet whatever future financial goal you have. Less volatile portfolios are almost equal relevant ingredients to be sure you get there.

What I am reading now:

Follow me on Twitter @nprotasoni