Image Source: Pixabay

The jumping off point for this post is the scene from A River Runs Through It when the little boy is writing an essay. He comes downstairs to show the essay to his father, played by Tom Skerrit. Skerrit reads it silently, marks it up a little, gives it back to the little boy, and says, "half as long." The little boy goes back upstairs, comes down a little later after he presumably shortened it by half. Skerrit again reads it in silence, marks it up a little, and then repeats "half as long."

I think there is a lesson in there that could be applied to portfolio complexity. "Half as complex." 

Larry Swedroe, writing at Alpha Architect, wrote about a study by Goulding and Harvey called Investment Base Pairs. It is some sort of strategy that offsets exposures, which could include long/short, but doesn't necessarily have to be long/short. The Swedroe write-up wasn't that clear, to be honest, so I asked Claud.ai to create a portfolio using mutual funds and ETFs to replicate the Investment Base Pairs Strategy.

I swapped in URTH because Claude suggested a version of that fund that only trades in Europe, and the weighting to BIL was split between two different cash proxies. The fourth portfolio is just VBAIX.

I don't know whether Claude had it right or not, but its portfolio certainly produced a valid result that outperformed VBAIX with a similar volatility profile but not much differentiation. The Slightly Simpler Version of Claude did quite a bit better due primarily to SPMO and QLEIX. It all blends together pretty well, and another contributing factor to its success versus VBAIX is that there are no bonds with duration. 

The Much Simpler Version of Claude was lights out. The CAGR was almost double the original Claude and almost triple VBAIX, all with a small fraction of the volatility, and it was up in 2022. I threw in the RISR/MBB pairing from yesterday. That was just coincidental timing between yesterday's post and today's post. So, should we just back up the truck on the Much Simpler Version of Claude?

The backtest is skewed by the insanely good performance of QLEIX. It compounded at 27% for the period studied. I would not want to bet that it could repeat that same sort of growth going forward. I think the volatility numbers could stand up, I think it could be a source of crisis alpha when needed more often than not, and I think the low correlation to the S&P 500 can persist, but relying on that sort of growth will likely end up in disappointment. Not tears, though. If it compounded at 1/3 that rate, I think the portfolio could be pretty robust in terms of more of an all-weather approach, and that is important. I would not have 40% relying on the RISR/MBB pair working, though. I would cut the exposure to that trade in terms of thinking about any sort of real-world application of this. 

The ReturnStacked guys launched the US Stocks & Gold/Bitcoin ETF (RSSX). "For every $1 invested, RSSX is designed to provide $1 of exposure to U.S. equities and $1 of a Gold/Bitcoin strategy." Digging in a little deeper, they are weighing the gold versus bitcoin by risk, in the neighborhood of a risk parity influenced strategy. Per testfol.io, it looks like RSSX would have 75% in gold and 25% in Bitcoin. Per Portfoliovisualizer, it would be more like 95% in gold, 5% in Bitcoin. As of now, the positions aren't listed, but I believe I am directionally correct. Tying in to Half As Long, it is hard to find the benefit to the added complexity of their funds, but we'll see about this one.

While we're on it, the ReturnStacked guys maintain a suite of model portfolios that use their funds, yes, but also funds from other providers. Let's revisit the All-Terrain 12% Volatility Target; the names are great. It may not be ok to list out the holdings and weightings, but they use a lot of their funds and some other levered funds. The asset allocation shows 28.4% in equities, 49.7% in fixed income, real assets at 39.5% and 44.3% in alternatives for a total of 161.9%. The weightings are added up on a look-through basis, so for example, a 20% weight in RSST would give the portfolio 20% to equities and 20% to alts (managed futures).

All-Terrain appears to be a permanent portfolio/quadrant style inspired by the following comparison, which is time-constrained by a couple of funds in All-Terrain.

The previous six months have been a great litmus test for all sorts of ideas, but not yet as useful as 2022. The shortness of the period studied does bring in some skews for how well gold did, and 25% in cash helped as the market has chopped around so violently. The Permanent Portfolio Mutual Fund (PRPFX) is more complicated than Portfolio 3, to be sure, but it is much less complicated than All-Terrain. The 25% allocated to long duration in Portfolio 3 creates some vulnerabilities that I think are better avoided, but I would try to make the simplicity of Portfolio 3 work much sooner than I'd use Portfolio 1. 

I threw Cambria Trinity in because it is quadrant-inspired, and the allocation fascinates me, but it always looks like a difficult fund to hold. 

More By This Author:

A Fund That Shorts ETF Inefficiencies
Put It All In.....Long Short?
Friday Night Volatility