Thank you for the article! By chance I was also playing around with a Meb Faber type of monthly ETF rotation system. But I chose a more differentiated portfolio of 11 ETFs:
• Equities: VTI, EFA, EEM
• Bonds: BND, BNDX
• Real Assets: DBA, DBB, DBE, DBP, VNQ, VNQI
I buy only the top-ranked ETF based on factoring the 10 months performance and volatility. I also aim for a volatility target and have a stop loss filter. I rotate the ETF on a monthly basis.
Here are the backtest results for the period of 2014-2022:
• CAGR: 19.2%
• Sharpe Ratio: 1.07%
• SPY Correlation: +0.24
• Drawdown: -18.0%
I like this simple strategy, its KPIs and especially the low correlation to the stock market.
Volatility targeting and stop-loss have only minimal effects. What seems to be critical is factoring in the volatility during the momentum look-back period.
Thank you. As usual: Great article! I have tested a lot monthly rotation systems and I often see, that the stats worsen, if you test it on daily bars. Same here: the max DD increases from 18 % to 24 %, the Ulcer Index increases from 0,8 to 4,3. At least for me, I find the daily stats more relevant, because I will definitely look more often into my positions than just once a month.
Thank you for the article! By chance I was also playing around with a Meb Faber type of monthly ETF rotation system. But I chose a more differentiated portfolio of 11 ETFs:
• Equities: VTI, EFA, EEM
• Bonds: BND, BNDX
• Real Assets: DBA, DBB, DBE, DBP, VNQ, VNQI
I buy only the top-ranked ETF based on factoring the 10 months performance and volatility. I also aim for a volatility target and have a stop loss filter. I rotate the ETF on a monthly basis.
Here are the backtest results for the period of 2014-2022:
• CAGR: 19.2%
• Sharpe Ratio: 1.07%
• SPY Correlation: +0.24
• Drawdown: -18.0%
I like this simple strategy, its KPIs and especially the low correlation to the stock market.
What does the expert think?
Hi Marco,
Thanks a lot for your valuable input!
May I ask how many ETFs you include in your portfolio every month?
Volatility targeting and stop-loss have only minimal effects. What seems to be critical is factoring in the volatility during the momentum look-back period.
Thanks for your input. I will have a look at your ideas in a few days.
Only the Top1. I backtested Top2, Top3 etc., which leads to a performance decrease.
Thank you. As usual: Great article! I have tested a lot monthly rotation systems and I often see, that the stats worsen, if you test it on daily bars. Same here: the max DD increases from 18 % to 24 %, the Ulcer Index increases from 0,8 to 4,3. At least for me, I find the daily stats more relevant, because I will definitely look more often into my positions than just once a month.
Hi, Thanks for your comment. It's hard to not look at your positions, but my experience is that the more often I look, the worse my results.
It's crucial to detach from money, but it's not easy....
Agreed. That's why I'm having everything 100 % automated. This helps me not to make stupid things, but the emotions are still there...