Dual Momentum Investing With Mutual Funds

Gary Antonacci’s  “Risk Premia Harvesting Through Dual Momentum” paper available on Optimal Momentum first piqued my interest in using absolute and relative momentum to invest in small groups of asset classes. Antonacci states that “relative momentum looks at price strength with respect to other assets, while absolute momentum looks for an asset’s own positive excess return over a given look back period.”

Using ETFReplay.com I created three portfolios to test the efficacy of a strategy similar to the one outlined at Optimal Momentum.  The portfolios were constructed using mutual funds (a new feature of ETFReplay) as opposed to ETFs, due to the longer trading histories of many mutual funds. My objective was to create three portfolios representing three unique asset classes.

The first portfolio is an equity portfolio consisting of the following:

Vanguard Emerging Markets – VEIEX

Vanguard Short-Term Treasury – VFISX

Vanguard Total Stock Market – VTSMX

The second portfolio is a bond portfolio consisting of the following:

T. Rowe Price Emerging Markets Bond – PREMX

Vanguard Long-Term Investment Grade – VWESX

Vanguard Short-Term Treasury – VFISX

The third portfolio is a real asset portfolio consisting of the following:

Vanguard Short-Term Treasury – VFISX

Vanguard Precious Metals & Mining – VGPMX

Vanguard REIT – VGSIX

The strategy for the backtests involves purchasing one mutual fund in each portfolio and equal weighting the three portfolios to create a “complete” portfolio. Purchases are determined by the one fund in each portfolio which has the highest trailing 6 month returns. The strategy rebalances each month, selling the current holding if it is no longer the top ranked fund in its portfolio and replacing it with the fund which has the highest momentum.

In other words, the complete portfolio equally weights the top one holding in the three portfolios outlined above. At most the complete portfolio will have three holdings – an equity, bond, and real asset. In some months, such as late 2008, VFISX has the highest 6 month trailing returns in all three portfolios; therefore, the complete portfolio could have less than three positions depending on market conditions.

You will notice a short-term treasury fund is in each portfolio. Combining absolute and relative momentum means a mutual fund outperforms a risk-free asset class such as cash in addition to outperforming all of the other asset classes in the portfolio Thus, for the purposes of this test I added a short-term treasury fund to each portfolio to represent a risk-free asset and to act as a comparison point for absolute momentum.

Personal Capital
Finance Advice for the Digital Age.
Smart, Free and Entirely Personal!

The benchmark for the tests is the Vanguard 60-40 Balanced Fund (VBINX) which offers a better representation of  a complete portfolio when compared to a pure equity benchmark like SPY.  The strategy detailed above has offered strong historical returns at comparable volatility and much lower drawdowns compared to a balanced 60/40 mutual fund. Also note the Sharpe Ratio in excess of 1:

The results above since 2003 are impressive. However, before we announce to the world our discover of the holy grail, the strategy has under-performed a 60/40 investment over the past two years:

Recent under-performance offers us a lesson – strategies can work well for years and then struggle for months, years or never again. Prolonged periods of relatively low returns will test the patience of a strategy’s followers.  Has the tide shifted for good on the strategy outlined above? Or going forward can we expect something similar to the results from 2003-2010?

32 thoughts on “Dual Momentum Investing With Mutual Funds”

  1. I think a lot has to do with what kind of portfolio you select. I don’t know many people who would be comfortable having on-third of their assets in emerging market equities or gold mining stocks. I’d be curious to see what things would have looked like using a more reasonable international developed market fund in place of emerging markets, as well as something less volatile, such as diversified physical commodities (I think PIMCO has a fund for this), in place of gold stocks.

    1. Hi John, The objective here wasn’t to recommend it as a complete portfolio solution, but to test the strategy in an isolated environment. So, an individual could dedicate a percentage of their personal investments to this type of strategy. ETF Replay only has a limited number of mutual funds in their database, which is why the Vanguard Metals and Mining fund was used as a proxy for physical commodities.

      1. Why did you use emerging markets (80% max drawdown) instead of developed foreign markets? Performance depends on assets selected as well as strategy, doesn’t it? Even so, the last 2 years you report as under performance is still up 27%.

        1. I was looking for higher beta equity markets with lower correlations to US equity markets than developed foreign markets. I will try to post a follow-up using developed foreign equities

    2. Hi,
      Using Relative strength and momentum can be a strong edge. Proper asset class selection is the key. There has been a lot of money flowing into the S&P in the last few years. A universe that includes one leveraged component will significantly boost the returns while the volatility and drawdowns can be kept to a minimum. ETF replay is a wonderful tool.

  2. Would like to see the result sliced and diced over a one or two year rolling performance compared to the benchmark fund. Then the apparent underperformance most recently can be compared more effectively. Also I would say that the underperformance recently is mostly thanks to the over performance of bonds on a relative basis.

  3. How would this strategy stand up against the new hedge fund replicator ETFs such as GURU and QMN and ALFA?

  4. To directly compare it to the 60/40 fund, just use a 60/40 weight of equities and bonds only. No need for commodities or assets. Gary doesn’t use commodities in his momentum indexes anyways, just stocks, bonds and REITs. He also uses high yield bonds. He doesn’t like emerging market stocks or bonds but I think that’s more for liquidity issues for large institutions. I think it’s fine to use for the regular investor. There’s a time stop of 1 month so the risk is minimalized.

    At the risk of data mining, using a look back period of both 6 and 3 months, along with etfreplay fantastic volatility filter, I found this system to produce swell results, including the past few years.

  5. Am I right in noticing that the current Dual Momentum Portfolio has four buckets instead of the original three? You may have already discussed this somewhere, but could you go over the rationale for this change?


  6. Alex – The three buckets were used in ETFReplay backtest results due to their site limiting tests to 3 portfolios.

    1. Thanks, Scott.

      So, do I understand correctly that you believe that using four buckets will produce better results and/or smaller drawdowns?

  7. Hi, on what day are you observing the roll? if beginning/end of the month, could you switch the roll date to the middle of the month (I see very different results then).

  8. Yes, that is hugely important with bond funds, thank you for clarifying. In essence “Dual Momentum” is nothing more than trading the equity curve, whether of a fund or of your own trading. It is a simple on/off switch and much the same results can be achieved by switching to cash/treasuries when the climate turns sour. It is something I have experimented on considerably in the past and formed an important feature of my ETFs momentum system in my 2008 book. I have recently been looking at the benefits/disadvantages of turning an actual trading system on/off (although of course the principal is identical and any difference one of semantics). Unfortunately of course there is no “true” answer”. Standard deviation is always likely to be lower using such an approach and average drawdowns smaller. But returns are certainly not always higher and certainly not for all periods. If anyone is interetsed I will be posting my findings on my website shortly which will compare my Quantech 20 stock Smart Beta System without “dual momentum” to the same system with an on/off switch.

      1. Sorry to reply so late. No, I don’t think I’ll be writing another book any time soon. Your criticisms in your kind review where well taken: In retrospect I should not have linked it so heavily to the actual software I used for my research and I am making amends for that on my website where I hope I am explaining things a little better. I very much enjoy this blog and have added a link to it on mine. Regards A

    1. Nelson Freeburg explored this concept marrying the Pankin sector strategy and Pentad as his regime filter or “on off” switch. He achieved spectacular results (albeit during a secular bull market) with relatively low drawdown. You can search for saturday seminars Nelson Freeburg (RIP) to find the audio and accompanying text.

  9. Hello,

    Thanks for this excellent blog. I would like to test the Dual Momentum strategy with a Canadian focus to see if the results translate. I was wondering how you implemented the strategy in ETFReplay, is there a default strategy for Dual Momentum or do you have to import specific trades?


    1. To do it with a Canadian Focus you may have to import trades, but their support may also have some ideas for you.

  10. Hi Scott

    Just studied Gary Antonacci’s book, Dual Momentum Investing. In it he highlights his Global Equities Momentum (GEM) approach – which I like a lot for it’s results and simplicity.
    Do you prefer this GEM approach, that only uses 2 asset classes(stocks/bonds) over the 3-4 asset class which he uses in his 2013 paper?

    1. I find his 2 asset class argument compelling, especially since it was the focus of his book. He makes strong academic arguments for both models.

Comments are closed.