Accelerating Dual Momentum Summary

Accelerating Dual Momentum Summary

July 19, 2018 3 By Jeremy


Rules-Based Strategy Summary: Accelerating Dual Momentum

There are a lot of rules-based investment strategies available on the internet.  Some of them are good, and some of them are downright awful.  While researching the Dual Momentum strategy, I came across one of the good ones.  It is an adaption of Dual Momentum, called  Accelerating Dual Momentum.

The author of the post and strategy is Steve Hanly. He has run a very robust backtest on the strategy and I personally have decided to invest part of my portfolio with it due to the rigor of the backtest and the result obtained.

The strategy uses the Dual Momentum strategy as a background but has made some strong improvements that have made the system even more robust.

I am always leery of “tweaks” to good systems, as those tweaks can be simply data mining or curve fitting.  It is easy to look back at a strategy and over optimize for things that will not actually lead to greater performance.  Luck may be a factor, or there may actually be no correlation between those tweaks and what drives returns or drawdown protection in the system.

However, Hanly conducted robust backtesting, and was able to run the system going back to 1998.  This has the benefit of seeing how the system performs in multiple phases of the market.  In addition, he was able to test for over fitting by running the system across multiple time periods.  This reduces the chance that the good results of the strategy are due to some anomaly that has no chance of repeating again in the future.

Overall, this strategy is very interesting and the backtest results are excellent.  I will review more how the system works below at a high-level, but please be sure to head over to the post to get the full details.

As with other portfolios on Robotic Investing, I track real-world implementation of the Accelerated Dual Momentum system. To see how the portfolio performs in real-time and the real world, check out all the blog posts tagged with Accelerated Dual Momentum category.

Rules-Based Portfolio Construction & Management

Like the 12% Solution, the Accelerated Dual Momentum portfolio is built using momentum as the main driver.  The system differs from both the 12% Solution and Dual Momentum in the choice of momentum periods and assets chosen.

Momentum Periods

Accelerated Dual Momentum uses a combination of momentum periods to determine the best performing asset.  Specifically, the momentum score is the sum of each asset’s 1 month, 3-month, and 6-month returns.

Dual Momentum uses 12-months.

Hanly’s rationale for using these three periods is that it takes into account the curve of the trend.  By combining all three periods, the investor is getting an indication of the short-term and medium term trends.  Here is a comment he made with respect to the three periods:

One of the major elements of this strategy is how it looks to find consistent uptrends that are accelerating upward

Hanly expresses that he feels Dual Momentum’s 12-month momentum period is too long.  He states that the consequence of a 12-month period is,

…being too late to the party, by the time the signal may pick up a trend it’s already stopped and potentially reversing.

Asset Choice

The other difference compared to Dual Momentum, and where the accelerated part of the strategy really comes into play, is the choice of international asset.  The system uses two equity choices: the S&P 500 and a global small-cap ex-US ETF.

Hanly chose a global small-cap ex-U.S. fund instead of a total world fund ex-U.S. as it is much less correlated to large U.S. stocks.  This seems to have provided strong performance gains for the system.

In terms of tactics for the system, the investor calculates the three returns for each of the three funds (the system also trades a long-term treasury fund).  The ETF with the best momentum score is purchased.  However, that equity ETF is only purchased if the momentum score is above zero.

If the equity ETF with the best score is below zero, then an ETF tracking long-term treasuries is purchased.  This process is completed each month.   The fund with the worst performance is replaced.

Hanly has a great image from his site that outlines how the system works:

Accelerated Dual Momentum Process

Research & Backtest Results

For more information and detailed research on the U.S. Sector Rotation strategy, here are some key links to review:

Hanly has his full backtest results posted on both of these above links.  Using mutual fund data going back to 1998, the results have been very positive.  Since there were mutual funds with data going that far back, it made it easier to test the strategy through a number of different market cycles.  Testing with newer ETFs only provides part of the picture.

That said, when I test the strategy here on Robotic Investing (and run it in my own portfolio), I will be using comparable ETFs.

Here are the results with the mutual funds going back to 1998:

Accelerated Dual Momentum Returns to 1998

Very strong performance.  You can see why it piqued my interest and pushed me to start tracking it here on Robotic Investing.  What I really wanted to know for myself though was how it performed compared to a straight-up Dual Momentum strategy.

Here are the results using Total US/Total International funds (VTSMX/VGTSX) with Total Bond (VBMFX) for the Dual Momentum and the funds referred to in the Hanly’s post for ACD.

Accelerating Dual Momentum clearly came out ahead with total value of ~$436,000 versus ~$70,000 on a $10,000 investment.

Accelerated Dual Momentum vs Dual Momentum Returns

The main caveat is that performance with the funds chosen for Accelerating Dual Momentum will differ slightly from comparable ETFs.  Hanly talks about that on the site, and it is a concern.  One fund in particular that Hanly tested with is more active than an index fund so that may skew the result slightly.  However, there seems to be only one period in the backtest that is driving that difference, so going forward it may wash out.

As with other portfolios on Robotic Investing, I track real-world implementation of the Accelerated Dual Momentum system. To see how the portfolio performs in real-time and the real world, check out all the blog posts tagged with Accelerated Dual Momentum category.

Disclaimer: The information provided on this site is for education purposes only. The author is not a registered financial adviser and the ideas discussed on the site are just trading analysis and not recommendations. Robotic Investing doesn’t endorse any of the comments that might appear on the discussion threads. There is no guarantee for those comments to be accurate. By reading this site you automatically agree that Robotic Investing is not responsible for any of your trading decisions. Remember not to risk money that you cannot afford to lose. and all its products are Copyright© by Robotic Investing and property of Robotic Investing. All Rights Reserved.

Featured Image: Skyline nocturno de Boston by Guillén Pérez