- This topic is empty.
-
In light of recent changes to my own retirement account structure, and others taking an interest, I thought I’d post some info and data on my research down the All-Weather rabbit hole which has been going on for about 8 – 10 months now.
The original concept of the All Weather strategy comes from the massive Bridgewater Hedge fund. You can read that full story HERE
Salem Abraham, once a pure trend following CTA, has also moved to the All Weather model with his Fortress Fund. You can read that story HERE.
And lastly, Tom Basso, also once a pure trend follower, has moved into the same style and has recently published his book All Weather Trader, although from what I hear, its quite a more convoluted way of doing things.
The portfolio is intended to be relatively simple and easy to implement, with a mix of stocks, bonds, commodities, and currencies or ‘alternates’. Alternates can be anything – it looks like Abraham uses his old CTA trend-following model, but you could use Bitcoin, a covered call ETF etc etc
The core of the All Weather portfolio is a combination of 30% stocks, 40% long-term bonds, 15% intermediate-term bonds, and 15% commodities. The specific allocation is meant to be relatively stable over time, which is achieved through regular rebalancing.
The return profile is moderate and stable return with low volatility and low drawdowns.
The following is an example that uses 5 ETFs, each of which has an equal allocation of 20%. Each gets rebalanced back to 20% at the end of the month (IBKR makes this very easy with their Rebalance Portfolio tool). This would be deemed very conservative, essentially a 60/40 portfolio.
This portfolio, as is, has a CAGR of 9.2% with a maxDD of -14.8%. A few tweaks to the portfolio weights could get a little more juice out of it.
An important aspect here is to be very careful with the selection of ETFs for two reasons:
1. Many ETFs have limited data
2. They open themselves up to curve fittingTo overcome #2 I have built my two strategies solely using the highest liquidity ETFs in each sector. I have now got two of these All Weather portfolios, the main one trades US ETFs and is what I’m using to manage my retirement account. The other is an Australian version that trades ASX ETFs and we’re essentially going to use that for Shartru advisors to put their clients into.
You can also improve the performance a little more by adding a simple timing tool, albeit that’s not the initial intention of the All Weather idea. However, it makes sense to me in light of major and sustained bear markets in these instruments, i.e. stocks in 2008 and bonds in 2022 (up until 2022, Bridgewater had a max drawdown of some 13%, which blew out to 28% because of the bonds rout)
My two versions both use a simple timing tool
US
CAGR: +19.3% maxDD -14.7%ASX
CAGR: +11.4% maxDD: -9.8%The code above is what I have been using, but you could also use standard rotation coding as well, especially if you had a larger universe of ETFs that you wanted to use.
Zach has been playing with an ASX version using the rotational code on all listed ASX ETFs. It holds 5 positions with a CAGR of 9.3% and maxDD of -10.6%
Anyway, food for thought and happy to answer any questions.
Here is another type, called the Dedalo Four strategy, albeit starting to move away from the All Weather concept. It’s essentially an 80/20 portfolio, so 80% growth assets and 20% defensive.
It trades both the Vanguard Total US market and Vanguard Total World markets, plus the Vanguard International Bond ETF. In this example, I have added a timing tool, C > MA(C,200), to enter and exit.
CAGR: 9.3% maxDD: -10.9% (note: I have brought this one back into AUD)
If anyone knows, I’d like to know how to create a annualised volatility metric into the results table for RT
Firstly, many thanks Nick for sharing, tis something I’ve been considering from an SMSF perspective as well…much appreciated.
With regard Australian ETF equivalents would anyone have come across a reasonable resource for matching, or not as the case may be, US to AU ETFs?
With regards the scripts above, I’ve just started to research and have identified the following as possible candidates to explore further…welcome any thoughts as to reasonableness.
- You must be logged in to reply to this topic.