- The author shows a simple strategy using 2 of the most liquid ETFs. It needs a minute every month.
- It is supported by research and recent history: a 10% annualized return with a bond-like risk.
- The return can be widely improved by leveraging or extending the concept to more sophisticated models.
This article describes a strategy mixing in equal weight two paired switching models on treasury bonds and stocks. I have described in a book something similar involving different pairs. Here a single pair is used: the Dow Jones Industrial Average and the 7-to-10-year treasury bonds. These assets may be invested through futures. Hereafter I will use ETFs: the SPDR Dow Jones Industrial Average ETF (DIA) and the iShares 7-10 Year Treasury Bond ETF (IEF).
The first model "A" is a classic momentum-based paired switching. It invests every month in the ETF having the highest increase in price last 3 months. For the convenience, I use a 4-week rotation (not exactly monthly) and a price momentum of 63 trading days (considering 21 trading days per month on average).
The second model "B" is a seasonal paired switching. Unlike other seasonal models I have written about, it invests in the stock index only the 4 best months of the year: March, April, November, December. It is in bonds the rest of the year.
As a consequence, the portfolio is fully invested all the time: it may be 100% in DIA, 100% in IEF, or half in each (rebalanced every 4 weeks).
A short theory
This strategy combines the two oldest, best documented, and steadiest market anomalies: momentum and seasonals. Momentum paired switching has been the subject of a couple of articles since 2011, but the idea is much older than that. It is a simplified implementation of tactical allocation and relative strength, two concepts used for decades. Seasonal patterns have statistical data for more than three centuries (article here).
Not only are both switching models separately supported by research, they also have a better chance to work when used together:
- When the trend is consistent with the season, they add their gains.
- When it is not, they hedge each other.
A 15-year history
The next table shows performances since 1999, with a 1% precision. The strategy and its components are compared with holding DIA, IEF, and an equal-weight portfolio of both, rebalanced monthly. Dividends are included and reinvested, a 0.1% rate is taken into account for trading costs.
*Past performance, real or simulated, is not a guarantee for the future.
The Double Switch is the only line in the table where the annualized return is higher than the volatility (standard deviation). Its maximum drawdown is about 11%. It shows a strong resilience in market downturns. Three years are in loss: 2000 (-3%), 2002 (-2%), 2005 (-2%). In 2008, it shows a gain of 7% whereas the Dow Jones lost almost 34%.
Double Switch Total Return in % (blue line = Dow Jones index)
A 10% annualized return is not so impressive, but a low risk in terms of drawdown and volatility makes this strategy a possible candidate for leveraging.
No software is needed to implement the DIA-IEF Double Switch. The momentum component can be checked in a minute per month, and it results in less than a trading order per month on average. It looks among the simplest, safest and laziest strategies based on research. It is also a simple implementation of a powerful idea: mixing momentum and seasonal strategies may increase significantly the risk-adjusted performance. My ETF core portfolio combines more sophisticated models with the same idea (article to come).