Dual Momentum: Easy Trend Investing

One form of trend investing I really love for many reasons is Dual Momentum Investing.

Dual Momentum Investing has shown great results, it is easy to administer, it is pervasive across many market conditions, and it really reduces the kind of portfolio volatility that makes many investors do stupid, stupid things.

It might be the near-perfect solution to many self-directed investor problems—particularly human psychology.

Dual Momentum Background

The Dual Momentum strategy was brought to the public by Canadian money manager Gary Antonacci. He wrote a detailed white paper on the strategy a few years back and then published a book on the strategy as well.

Gary's Optimal Momentum website explains the strategy in good detail and his white papers are posted for your reading pleasure on SSRN.

The great thing about Gary is that he freely shares this strategy which is simple and has handily outperformed the broader stock market over whole market cycles (top-to-top or bottom-to-bottom).

Hearing several podcasts featuring Gary and reading his blog faithfully, I genuinely get the impression that Gary is an all-around good guy who shares this great information because he cares to make the investment returns of amateur self-directed investors better.

Gary is also super blunt. When criticized by many over the pervasiveness of the strategy over other investment approaches like buy-and-hold, his answer is simple: I'll be the first to tell everyone to ditch the strategy when it actually underperforms a whole market cycle.

Dual Momentum Results

Based on Gary's data—which is done in U.S. dollar investing—the Dual Momentum strategy has returned over 17% compounded annually since 1974. That beat the MSCI World IMI Index by approximately 9% per year over 4 decades!

Putting that into real money perspective, if you follow my minimum savings rules and put away $562 a month for 35 years, a regular buy-and-hold strategy would net you about $800,000 in today's dollars at 6% net return.

With Dual Momentum, if historical returns stay the same, you would have $8.2 million. That's a whopping 10x difference: the real power of compounding interest!

Even more appealing to me is this strategy has not had a draw-down of more than 20% since 1974. To get a comparable max draw-down with buy-and-hold investing, your portfolio would have consisted of more than 60% bonds. That would have left just 40% for the higher growth stock allocation. Your gross returns on a portfolio with similar risk would have been 7% per year (4.5% after inflation).

A lot of investors can tolerate a 20% loss without too much panic. Sure losing $100,000 on a $500,000 portfolio stings, but you can probably carry on with your plan.

Losing half your portfolio is what makes people go squirrelly, panic, dump everything into GICs, and moan and bitch to everyone who lends them an ear about the dangers of stocks and why their house is a so much better "investment".

Basics of Dual Momentum

Dual Momentum Investing is based on considering two factors: relative momentum and absolute momentum. Dual Momentum (GEM) strategy evaluates just three asset classes: U.S. stocks, international stocks, and cash using three different ETFs.

Basically you consider momentum factors and then invest 100% of your portfolio in the asset which has performed the best over a specified period of time.

Historically the invested asset class has changed just over once per year. This means minimal trading costs and little stress about picking the right day to evaluate. It really doesn't matter whether you are religious about checking in on your portfolio the first trading day of each month, or if you let it slip a few days every now and then.

For Canadians, I would suggest the best options for evaluation today are: XUU.TO for U.S. stocks, XEF.TO for international stocks, and XSB.TO for short-term bonds.

Once a month we evaluate our options and choose the best performing ETF for our portfolio holding. It's extremely simple and very effective.

For those who are okay doing something just a bit different but don't want to be a dedicated trader, Dual Momentum Investing may be the best long-term sustainable choice.

Evaluating Relative Momentum

Relative Momentum means comparing the historical performance of an asset with others in its broader asset class over a specified time period.

While opinion differs on the timeline that should be used, I personally like looking at the past 12 month performance and the past 6 month performance and averaging the two.

This is where we evaluate the total return performance of the two broad stock asset classes: XUU.TO and XEF.TO.

Here is the comparison on July 31, 2017:

XUU.TO 6 month: +4.13%
XUU.TO 12 month: +11.11%
XUU.TO Average: +7.62%

XEF.TO 6 month: +9.54%
XEF.TO 12 month: +13.22%
XEF.TO Average: +11.38%

By this relative momentum evaluation, XEF.TO is the clear winner. Therefore XEF.TO is our choice for the stock asset class.

However, we are not done here. Now we need to evaluate the performance for the absolute momentum side.

Evaluating Absolute Momentum

Absolute Momentum is the evaluation of an investable asset's performance compared to cash (T-bills) over a specified time period. Again, I like averaging the 12 month and 6 month performance.

T-bills (very short-term government bonds) are used here because they are widely perceived to be a nearly risk-free asset class. A short-term bond index ETF (such as VSB.TO or XSB.TO), does a decent job of substituting for T-bills in this evaluation.

Some other writers, including Antonacci himself, compare to T-bills but choose a broad bond index for their investment choice. To me, using a short-term bond index ETF for our comparison and investment choice is easier and just as effective.

Here is the comparison as of July 31, 2017:

XEF.TO 6 month: +9.54%
XEF.TO 12 month: +13.22%
XEF.TO Average: +11.38%

VSB.TO 6 month: -0.31%
VSB.TO 12 month: -0.59%
VSB.TO Average: -0.45%

In this absolute momentum evaluation, XEF.TO is again the clear winner.

So, currently you would have 100% of your portfolio invested in just one ETF: XEF.TO. This gives you an indirect stake in 2,500 companies located across Europe, Australia, and Japan.

I will be publishing the Dual Momentum recommended signal based on this evaluation every month in the Portfolio Update posts. I think it would be a great strategy to keep an eye on.

Comments & Questions

All comments are moderated before being posted for public viewing. Please don't send in multiple comments if yours doesn't appear right away. It can take up to 24 hours before comments are posted.

Comments containing links or "trolling" will not be posted. Comments with profane language or those which reveal personal information will be edited by moderator.

33 Replies to “Dual Momentum: Easy Trend Investing”

  1. Hi Daren,

    Can you execute the Dual Momentum Strategy in the TFSA/RRSP accounts? Ideally I’d like to keep the potential profits tax free ( or defer the tax) but I would hate for the CRA to come knocking at my door. What’s your take on this?

    Thanks.

    1. Mr. Rich Moose says:

      I’m not an accountant, but I don’t think it would be a problem. High trading frequency can be one of the factors to determine if the CRA considers you to be operating a “trading business” which is not allowed in a registered account. However, averaging less than two trades per year and only one required account log-in per month, DM is far from a higher frequency trading strategy.
      I’m not aware of any cases related to this issue that deal with DM specifically. The court case that set the “trading business” standards seemed to be a guy that traded quite frequently, used leverage, used options, was a professional CFA, and turned over positions at predetermined profit levels. I personally consider it to be a strange case with a curious strategy. An interesting read though: Foote vs. The Queen.

    2. Yanniel says:

      This is my understanding as well Daren, but since you are executing DM I wanted to know if you are doing so in a registered or non-registered account. I realize this question is private; so, please, ignore me on this topic if you feel I am asking to much of you.

    3. Mr. Rich Moose says:

      I’m a little hesitant to share exactly what I invest in at this point as my strategy is quite aggressive. I have family and friends who read the blog but don’t share the same tolerance for risk that I have.

    4. Yanniel says:

      I understand. Thanks.

  2. Hi Daren,

    Another comment containing a few considerations and questions, please. I know this is a lot, but when you can, I would greatly appreciate your input.

    One:

    Let’s say you want to make new deposits to an account executing the Dual Momentum Strategy. At which point will you buy into the trend? Will you wait to pick your ETF winner for the month and at that moment buy the winner with the new cash? In this scenario there’s an ETF already bought and we are bringing new external cash in.

    Two:

    I have reservations about using VSB as a substitute for T-Bills without running a backtest. By any chance did you back-tested this substitution? VSB holds mainly Canadian bonds, not American. Although the BoC follows the FED in lockstep ~90% of the time, the BoC can delay several months to follow the FED’s lead. Because of this VSB and T-Bill could differ enough on a monthly basis to affect the winner of the Dual Momentum for the month.

    Also, VSB has longer durations than T-Bills (~2.xx vs 1 year). On top of that, ~25% of VSB is allocated to corporate bonds. I would assume T-Bills were used because they are the safest. VSB although safe, is not the safest given that it has longer durations and exposure to corporates. This difference might cause you to pick a wrong winning in the Dual Momentum Strategy.

    About safeness again, the US bonds are perceived to be safer than the Canadian bonds. Also, having your investments in US dollars (as opposed to CAD) can act as a shock absorbent when the downturns happen.

    All these differences makes me wonder if VSB is a right replacement for T-Bills.

    Three:

    The equity ETFS you are using are unhedged, right? If so, have you thought about the implication of currency fluctuations? The underlying indexes can render positive results, yet the ETF could perform poorly or negatively (or better for that matter) because it is been tracked in CAD. The forex movement adds an extra variable and again I don’t know if these Canadian ETFs would be nice substitutes to the ones back-tested by Gary.

    Four:

    In Gary’s back test, did he use an equity index for the whole developed market except US? If so, XEF is not the same. Just as an example, XEF does not contain Canadian equity. I just worry that the ETFs that you are using would render completely different results than the ones back tested by Gary.

    Thanks!
    Yanniel.

    1. Mr. Rich Moose says:

      Yanniel, I’ll try my best. Unfortunately free data from Canadian sources is rare, so I have to use what I can…
      1) There’s no wrong answer here and it wouldn’t make a difference over a long time period, but I would add to the existing position.
      Let’s separate this into two issues: a) mixed short bonds vs T-bills, and b) currency.
      2-a) There’s definitely a valid argument to short bonds vs T-bills. T-bills are the ultimate in security as a government default on short-term bonds is very, very unlikely. Corporate shorts have a slightly higher default rate. Given that corporate shorts are 30% of VSB, the drawdowns on VSB will be higher than T-bills alone. However, even during the financial crisis corporate shorts had a maximum drawdown of 7.6%. Since 1992, they’ve outperformed T-bills by 1% per year quite consistently. In his evaluations, Gary uses T-bills as the hurdle but actually invests in the U.S. Aggregate Bond market (Canadian equivalent is XBB.TO/VAB.TO). If VSB.TO is a concern for the hurdle, you can use the risk free rate of Canadian T-bills by pulling Bank of Canada data. However, I would still invest in VSB.TO/XSB.TO or even an broad bond ETF if the signal says to go in bonds. ZFS.TO would be our closest ETF option to T-bills.
      2-b) Currency and bonds is important. It’s rarely a good idea to invest in foreign bonds without currency hedging to your home currency as it adds unnecessary volatility. https://www.vanguard.com/pdf/icrifi.pdf
      The point of the bond allocation is capital preservation, not trying to make as much as possible off of volatile currency swings. Gogi Grewal did a nice back-test on this. His finding say converting currency and buying stock ETFs on U.S. exchanges but bonds on local exchanges is best. This is followed by all U.S. exchange products. The difference between the two approaches is very minor. Interestingly the drawdowns using local currency bonds is lower and the total returns are higher.
      I would differ with the view that U.S. T-bills are safer than Canadian T-bills. Currently U.S. debt is cheaper than Canadian government debt by about 20 bps. Maybe something to do with our poorer stock performance, maybe its the fact that our federal government debt is much lower than the U.S..
      3) Yes, unhedged products. I don’t think there is a real wrong answer here since hedged products are getting quite good at keeping their hedging costs very efficient. Having an unhedged product could be considered very similar to converting your currency to USD and buying a U.S. listed ETF. I.e. if you buy XUU.TO, you will get almost the exact same results as converting to USD, buying VTI, and converting the returns back to CAD.
      Of course the results are always going to be different from Gary’s results, but that’s always an issue when you earn and save in a non-USD currency. You may sacrifice a couple percent in one five-year period and make it back the next period, but that’s the nature of investing with more than one currency in the picture. For ease of transactions, I think there is not really anything wrong with choosing local, low-cost ETFs for most people. (Large accounts might be a different story).
      4) Yeah the EAFE vs ACWI-ex US is an issue. EAFE results have been lower since 1988. (Gary’s back-tests actually use EAFE prior to 1988.) However, if you want to use local exchange products, EAFE is the best we have as there are no ACWI-ex US funds available to us. If you want to see some really interesting results, plug Emerging Markets only into your back-test from 1988 onwards. Dirty secret: It actually significantly outperforms standard DM (GEM).

  3. Hi Daren, I tried yesterday to post a comment containing links? Are links forbidden? I don’t see my comment, not even in moderation stage. Let me know what the rules for commenters are if you think I am breaking any. Thanks.

    1. Mr. Rich Moose says:

      I’m not allowing any comments which contain links. (It’s a way to prevent spamming or promotion of low quality websites/blogs).
      If you like, you can describe your link and I will find it, verify the content, and post the link so others can read.

    2. Yanniel says:

      I see. That’s ok.

      I was referring to some articles about ETF rotation written by “The Lazy Trader”. I found those to be very interesting and relevant to the general topic of momentum investing.

      Question: do you know a good (online) system to back-test Canadian ETFs?

      Thanks.

    3. Mr. Rich Moose says:

      For current ETFs Portfolio Visualizer is a great tool. However, the parameters are limited.
      If you’re looking for more historical data, it’s MSCI and S&P Index data and spreadsheets.

    4. Yanniel says:

      Thanks for the tip.

  4. Mr. Rich Moose says: Reply

    For anyone interested in reading, The Lazy Trader has some great content. I believe this is the article series Yanniel is referring to: http://www.the-lazy-trader.com/2015/01/etf-rotation-systems-to-beat-market-global-equities.html

  5. Daren, do you think it makes sense to combine leverage with this strategy? I mean “conservative” leverage; nothing too greedy. If so, do you think the margin interest (or loan interest) would be deductible in the eyes of the CRA?

    Also, what about using swap based ETFs with dual momentum instead of the traditional ETFs you are using? My thinking is that with your current ETFs you are must pay tax on the dividend income. But with swap based ETFs you would not pay tax on dividends, only on capital gains (if any).

    1. Mr. Rich Moose says:

      Careful use of leverage (1.25:1 or 1.5:1) will provide a nice boost to returns over the long run. Too much leverage will increase the risk, max drawdowns, and cost of whip-saw trades.
      If you follow DMI in a non-registered account and buy conventional ETFs, the interest costs will be deductible. (You will need to pay tax on the investment income too.)

    2. I was actually thiking even a more conservative leverage: 1.18:1. (Ex. I control 100k of assets by putting down 85k). I came out with this number after reading an article on Investopedia: “Finding Your Margin Investment Sweet Spot”.

      I appreciate your input!

  6. Other points about using swap based ETFs:

    1.) it is not just the taxation of the dividend that worries me; but also the taxation of the interest income when holding the bonds ETF. The latter is taxed at our individual marginal tax rate.

    2.) swap based ETFs track their indixes very well; which is a plus.

    Another thing that crossed my mind. If one were to use margin with swap based ETFs, I don’t think the margin interest would be tax deductible. I think there’s a requirement that for the margin interest to be deductible, the assets must produce divident/interest. Swap based ETFs don’t pay either.

    1. Mr. Rich Moose says:

      Swap ETFs are good for U.S. exposure (HXS.TO) and bond exposure (HBB.TO), but not for international. So you would still need to buy XEF.TO (or similar) when the signal calls for it.
      You’re definitely correct about borrowing to buy swap-based ETFs. Don’t count on the interest to be tax-deductible.
      Another consideration: The U.S. market offers a lot more swap and leveraged ETFs. Direxion and Proshares are two big providers. It could be worth trading in USD just to acces those funds. Portfolio+ (owned by Direxion) has some nice 1.25x ETFs.

    2. Thanks for yor insights Daren.

      I have my reservations about the leverage ETFs. They rebalance at the end of each day, don’t they? This will mostlikely cause the ETF to underperform in the long run. I think?

    3. Mr. Rich Moose says:

      Yes, most of these leveraged ETFs are daily re-balancing. This means they under-perform in high volatility environments and outperform in low volatility environments.
      They are not a buy-and-hold instrument because in down markets they do 1.25x or 2x or 3x as bad as the index. Over an entire market cycle, they will certainly perform worse than the index. But I think they can be used effectively for SOME people in a controlled timing strategy.
      I’ve used leveraged ETFs for some time now and have been happy. They perform as I would expect. 2x each day and a little less than 2x in an up week/month/quarter/year after fees.

    4. Yanniel says:

      Thanks Daren,

      Have you written anything about strategies using these leverage ETFs.

  7. Mr. Rich Moose says: Reply

    I’ve got a post just about ready that should go up Tuesday actually. It’s based on a back-test I’ve done with the U.S. index going back to 1970. If you use leverage, or leveraged ETFs correctly with enough safety, it shows some pretty amazing results!

    1. Yanniel says:

      I’d love to read that!

  8. Hi Daren,

    My gut feeling was telling me XEF wouldn’t be a nice ETF for INT EQ. I found some reasearch on the topic on Gary’s blog. The culprit for XEF’s underperformance seems to be the lack of Canadian+EM exposure. In particular the lack of EM worries me, because as time passes the growth of the world will be in EM.

    I am playing with the idea of creating “my own ETF” to bundle in XEF, XEC and XIC. There are systems that would allow you to do such things. But I have to investigate more on this. Otherwise, I think I am going to do my trades on the US market. I’ll probably do Norbet’s Ganbit to exchange my CADs for USDs.

    Also, this same article considers in great detail currency fluctuation implications for non-us investors when implementing GEM. It even cover the topic of currency hedging.

    Finally, the article concluded that it’s better to perform the price analysis in USD even when you do the trades in CADs. I think you can benefit from this.

    Name of the article “Dual Momentum for non-US Investors”.

    1. Mr. Rich Moose says:

      http://www.dualmomentum.net/2015/06/dual-momentum-for-non-us-investors.html
      Great read for anyone interested in DM strategies. In fact, read every article, white paper, and book by Gary plus listen to any podcasts he’s been on. It’s all fantastic stuff and important to understanding the principles of the strategy.
      After having read all of his stuff, my belief is adding Canadian & Emerging stocks to a Canadian dollar account is pointless unless your account is quite sizeable.
      For pespective, Gary did not use any Emerging markets in his models pre-1988. Dual Momentum performed quite nicely from 1970-1987.

    2. Yanniel says:

      Hi Daren,

      I would agree that the Canadian exposure could be ignored since it’s not significant now and it’s future growth won’t be significant either.

      It is very different for Emerging Markets. EM were not as important in the past as they are now. Their influence before 1988 is not as noticeable as today’s. And this influence will keep growing. Consider this:

      “emerging markets may produce a larger % of world GDP: from 45% currently, to possibly 70% in under 3 decades.”

      I took the above from an article covering the importance of EM in DM. Should you want to read it google for “What If Emerging Markets Eat Our Lunch?” by Gogi Grewal.

    3. Mr. Rich Moose says:

      For the purposes of my model, I’ll be comparing U.S. and Developed stocks because it’s the simplest for what I’m trying to do. That is, share a simple Dual Momentum model with Canadians using Canadian-listed products.
      I’m not saying you can’t invest in Emerging markets. If an investor believes the out-performance is there, go for it.
      That said, I don’t buy Emerging market hype or stories. GDP growth does not equal stock market investor returns. I try to ignore nice stories and focus on price and simplicity. It makes for clearer thinking and better long-term investing.
      Gogi’s article: https://growresearch.com/blogs/blog/what-if-emerging-markets-eat-our-lunch

    4. I hear you Daren. I really do. I get what you are traying to do by keeping EM out. The reality as Canadians is that we don’t have an Equity ex-US ETF that includes the developed world plus EM. By using XEF; you are simplifying the model and for you that is acceptable.

      That is different however in the American market. It is very easy to buy VEU; a single ETF covering the developed world (including Canada) plus EM. I favor this view, but I respect yours; which is simpler to implement.

      To be clear: I was not suggesting to add a 3rd ETF (something like XEC) for the purpose of executing relative momentum on EM as well. My suggestion was simply to include EM proportionally as part as our INT EQ assets.

  9. Hi Daren,

    One observation: The returns of your DM from 2014-2017 are quite different from the GEM returns posted on Gary’s blog. Yours are better I think. This makes me think something is fundamentally different in both implementations and that maybe you got lucky with your better results. Any theories?

    1. Yanniel says:

      Actually, part of the difference could be explained because the CAD was in a downtrend since before 2014 all the way to 2016. Since you use unhedged ETFs that was good for you.

      Your look-back period is also different. What’s your rationally in using in not using the recommended 12 months look back?

      Other difference is your lack of exposure to CAD EQ/EM EQ.

    2. Yanniel says:

      Another difference could be the use of leverage on your part.

    3. Mr. Rich Moose says:

      There’s no leverage in my DM models. I suspect the difference is 2 factors: currency and the slight difference in timing. Currency being the bigger one.
      The look-back is slightly different based on the Jegadeesh & Titman papers. They found trend to show outperformance after 3-12 months. By averaging 6-month and 12-month, the model becomes slightly more responsive in reducing drawdowns. The trading frequency increases a bit (+20%), but should still be less than 2 trades per year on average.

    4. Yanniel says:

      Thanks for your take.

Leave a Reply