Dual Momentum: A Year in Review

1 Flares Twitter 1 Facebook 0 Filament.io 1 Flares ×

A little over year ago I took a drastic step. I moved all of my retirement account holdings away from the safety of a conventional buy-and-hold passive index portfolio, and towards a different approach; an approach riddled with dangerous taboos.

  • It was an active approach.
  • It was market timing.
  • It was dual momentum investing, more specifically Global Equities Momentum (GEM).

And I wrote about it. I wrote about my reservations, my motivations, the knowns, the unknowns, why I thought it was likely to work for me, and why it was almost sure to be simple but not easy for me to stick to the plan.

And since that time, a lot has happened.

  • I started a successful online business, with my friend Brad Barrett, systematically teaching people how to play the miles game from square one.
  • My medical practice became evermore busy, with weekdays off increasingly becoming a thing of the past.
  • I tangled with the moderator over at Mr. Money Mustache’s forum, (not to be confused with Pete, Mr. Money Mustache himself, who is truly a stand up guy) about dual momentum one too many times, and got myself booted from the forum for good in an ugly display of moderator overreach.

Which is all to say: I’ve been busy.

In other words, I haven’t done a lot of writing here about my actual experience investing this way since I took the leap.

And now that it’s been more than a year, I figure it’s time for an update.

Where to begin? Well if you’re anything like me and you are a performance chasing fool, you’ll want to know how the whole thing worked out. Did dual momentum beat the market? Did it diminish drawdowns?

So before I get to my analysis of the experience, I’ll skip to the punchline.

It didn’t beat the market.


(The Horror…the horror…)

How it went down…

I began this experiment at the beginning of September 2014 when I moved over all of my retirement holdings into VIIIX, vanguard’s institutional S&P 500 fund.

And there it stayed until May 1, 2015.

Over those 9 months VIIIX Returned 5.54%.

At the end of April I got my first signal to make a trade. For the six months preceding May 1, FSPNX (fidelity’s International developed market index fund) had outperformed both short-term treasuries, and the S&P 500. So prior to open on April 1 I moved my holdings to FSPNX, and there they remained until September 1, 2015.

Over those five months FSPNX returned -8.4%. (Greece happened). During that same time period, The S&P 500 returned -4.7%.

And then in August we had our first big correction, and the markets entered an extremely volatile and negative period, giving both international developed markets and the S&P 500 a negative total return in the prior six months relative to short term treasuries.

And so before open on September 1, I moved all of my shares of FSPNX into VBITX (Vanguard’s short-term bond fund.)

And for the two months ending October 31, 2015, VBITX returned 0.11%. During that same time period, The S&P 500 returned 6.2%.

So all in all in 14 months of investing my nest egg using the dual momentum playbook, my retirement accounts returned – 2.8%.

And there are number of things that this result should tell you.

  1. If I had launched a dual momentum fund at the same time that I started investing my own nest egg in this manner, it’s safe to say that big money would not be flowing into my fund right about now.
  2. Dual momentum underperformed the S&P 500 by 8.2% (a very significant amount) over this time period.
  3. During my first two months in short term treasuries the S&P 500 ripped off an impressive rebound from its August lows, returning 5.7%. (I’m still in VBITX, so this chapter isn’t finished yet, but at this point, as I write this, it sure feels like the market volatility has passed for now. ) It was this flight to safety that caused the vast majority of my investing period’s underperformance.
  4. Of course I was not invested in 100% VIIIX, prior to adopting dual momentum, so my personal benchmark is not the S&P. My benchmark is what I was invested in before adopting this strategy. And luckily I wrote about my buy and hold portfolio before ever embarking on this journey. That portfolio returned -2.64% compared to my dual momentum portfolio’s return of -2.8% percent for this 14 month time period.

So the bullet point results for the past 14 months:

  • Global Equities Momentum significantly underperformed the S&P 500.
  • Global Equities Momentum displayed no significant difference in terms of performance from my previous buy and hold small value tilted internationally diversified portfolio.
  • Global Equities Momentum’s underperformance was largely attributable to its 2 months in short term treasuries (thus far) during a bounce back of the stock market.

Now obviously in any active strategy, successful or not, there will be periods of negative tracking error during which time the strategy will significantly underperform the market.  This is simply the cost of doing business for any american investor who decides not to invest 100% of his portfolio in the S&P.

And obviously the 14 month look back period is far too short of a timeframe to come to any meaningful conclusion about the odds of success for dual momentum as a long term investment strategy.

But the truth is that actually investing in such a strategy is experientially very different  from backtesting a strategy. And feeling the undulations of performance in real-time of course raised some new questions.

It is these questions that I really wish to adress in this post.

Question number one: How unusual is it for global equities momentum to NOT outperform the S&P 500 on a yearly basis?****

Answer: more common than a coin flip coming up heads.

To answer this question I plugged in DFALX (foreign developed), VFINX (S&P 500), and cash into the portfoliovisualizer.com dual momentum back-tester, and examined annual returns from 1993 to present (The longest available period for review when including an index fund of international developed equities.)

The results were surprisingly balanced.

  • Global equity momentum beat the S&P 500 for 10 of the 23 years.
  • Global equity momentum lost to the S&P 500 for 10 of the 23 years.
  • Global equity momentum and the S&P 500 had identical returns for three of the 23 years.
  • So the odds of global equities momentum losing to the market on any given year was 43%, and the odds of losing or tying the market was 57%.
  • And in losing years there was a wide range of possibilities in terms of the degree of global equities momentum underperformance.(-0.72% to -25.02%). Viewed in this light, an underperformance of 8% is hardly unusual, and a future underperformance of at least 25% is to be expected.
  • The longest continuous period of underperformance was six years (going on seven) from 2009 until 2014.

Question number 2: How unusual is it for Global equities momentum to underperform the market while in cash? (This is a very important question I think, because a principal feature of the dual momentum strategy’s past success has been its ability to avoid big drawdowns during bear markets.)

Answer: it is extremely common.

  • In looking back to 1993 there were 14 different timing periods where the dual momentum global equities momentum portfolio held cash as an asset. During those 14 different Holding periods, The S&P 500 outperformed global equities momentum 10 times.
  • So if this is the base rate scenario, then a global equities momentum investor (like me) should expect with 71% certainty, that he will underperform the market each time he exits the market.
  • And make no mistake about it, this underperformance can be significant and painful. Following the flash crash of 1998 global equities momentum underperformed the S&P 500 by 21% while in cash.

Question number 3: For those periods when the dual momentum investor held foreign equities instead of the S&P, how common was it for global equities momentum to underperform the S&P?

Answer: it was barely more common than a coin flip.

  • In our back test there were 17 periods where the dual momentum investor held foreign equities.
  • In nine of those 17 periods The S&P beat foreign equities in total returns.
  • In 8 of those 17 periods foreign equities beat the S&P.
  • The worst period of underperformance was May 1993 to November 1993 when global equities momentum was invested in foreign equities which underperformed the S&P 500 by 5.88%.

So let’s review what we have learned today.

If history holds, then as a global equities momentum investor trading on a six-month look back period , I should expect to…

  • Underperform or tie the S&P 500 more than half the time on a yearly basis.
  • Underperform for a single calendar year of returns by at least as much as 25%.
  • Underperform the S&P 500 for at least six years in a row.
  • Underperform the S&P 500 upwards of 70% of the time while holding cash.
  • Underperform the S&P 500 about half of the time while holding international equities.

Which is precisely why this strategy, while simple, will never be easy.

This is a strategy which performs at its best during big bear markets. During those time periods I will still be losing money (particularly in my taxable accounts), just not as much. So my wins may not feel so sweet.

More than half the time I will feel I am losing to the market, because I will be.

And there’s no reason why there cannot be long stretches in the future where I will objectively be losing to the market year after year.

Those are the facts.

But I’m still very bullish on the strategy, and on my ability to stick with it through thick and thin.

This is because I have two fundamental beliefs when it comes to investing.

  • An investor should keep it cheap.
  • An investor should focus most of his energy on minimizing downside risk.  (ie the permanent loss of capital.)

It in my view dual momentum has done a great job of achieving both of these goals. Particularly number two.  And I see no reason why the future performance of this strategy should be any different.

But in the end, the proof will be in the pudding.

Screenshot 2015-11-15 14.38.07

(Yesterday’s pudding)

***Note My results actually correspond to a timeframe of 14 months duration , not a year . Dual momentum looked better at the end of August which was the end of my first year investing this way. I chose this longer time period because it contained the most available data (even though it made GEM performance look considerably worse.)

The 12 month total returns for those interested from September 1, 2014 to September 1 2015, were….

  • GEM:   -2.86%
  • VIIIX:    0.4%
  • My prior portfolio:   -4.84%



1 Flares Twitter 1 Facebook 0 Filament.io 1 Flares ×

30 Responses to “Dual Momentum: A Year in Review”

  1. Robert K November 15, 2015 at 4:50 pm #

    I am a bit confused and while my GEM portfolio also is off compared to SPY, I think that my rules may be different from yours. The classic GEM, as described by Antonacci, uses a 12 month look back, and if I read this correctly, you are using a six month look back.

    On October 1, the 12 month look back gave the “sell” signal, moving me from SPY (or IVV, in the Fidelity environment) to fixed income. Of course, we know now that the market went up 8.4% in October, giving us a “buy” SPY signal on November 1 and giving me a sick feeling in the stomach.

    I reached out to Gary Antonacci on Twitter inquiring how often this occurs, and he indicated that it was not all that uncommon. He referred me to his blogs for more information and a quick survey failed to reveal any additional discussion.

    Disclosure: The GEM comprises about 15% of my overall investment portfolio but guided new money investment decisions. I also have smaller allocations for socially responsible DM and fixed income DM.

    • Miles Dividend M.D. November 15, 2015 at 8:03 pm #

      Yeah, Robert. Gary definitely advocates for the 12 month look back period, and for good reasons.

      -It is less susceptible to data mining since this was the original period described for the momentum anomaly.
      -It decreases the frequency of trading

      When I chose the six month time period my thinking was that it would get me out bear markets quicker and into bull markets a little sooner, so it would be behaviorally easier. now I think the best idea is to split your portfolio into buckets with different lookback periods to avoid 100% exposure to period specific whipsaws, when possible. (I do use 12 month lookbacks in my IRA/HSA accounts). But if I had to choose one lookback only, I would now probably choose 12 months as Gary recommends, because less trading will always be cheaper in terms of friction.

      Interesting that you got back into stocks before me, short treasuries actually outperformed the S&P for September, but you had a couple percentage points of better performance may-september, when I was in foreign developed, so you are still ahead.

      One thing that was really interesting to me, is that even with 75% of my portfolio in short term treasuries, it was still the performance of the stock market that determined the daily performance of my portfolio. Stocks just move so much more.


      Thanks for the comment, and good luck,


      • Fred C Dobbs November 16, 2015 at 8:45 am #

        On the FAQ page of his website, Gary indicates that shorter look backs lead to lower returns, more whipsaws, and higher drawdowns over time. I prefer to stick with 12 months as he does.

        • Miles Dividend M.D. November 16, 2015 at 9:01 am #

          Right. He does. And I think Gary’s prbably right, mostly because of the trend towards decreased trading costs.

          On the other hand Steve lecompte of CXO argues for a 5 month lookback.

          When you backtest multiple lookback periods for the same DM stategy, there is no clear winner across different asset pairings, but the shorter the lookback the more there is a tendency to trade, and all things being equal, trading is a source of expense which will eat away at returns.

          I actually favor splitting one’s portfolio into buckets with different lookback periods (say 3,6,12 months) to diversify away period specific whipsaws. And I do use a 12 month lookback in my non employer based accounts (HSA/IRA).


          • Fred C Dobbs November 16, 2015 at 9:15 am #

            I don’t think it’s just trading costs that cause shorter look backs to under perform. There are also more whipsaws and lost opportunities for profit. Gary emailed me that using a 3 month look back gives a larger worst drawdown than buy and hold.

          • Miles Dividend M.D. November 16, 2015 at 9:26 am #

            Its sporadic. Pick some asset pairings and do your own 3,6,12 month lookback period back tests. You wont find that 12 months consistently wins in terms of returns, draw downs, risk adjusted returns, more than any other lookback period. But you will find that it has tends to have less trades, and thus trading costs.

            My guess is that with shorter lookbacks you will have more whipsaws, because you will have more trades, and thus more opportunities for whipsaws to randomly occur. But I doubt you will have a higher percentage of whipsaws per trade since whipsaws are likely random events.

            Just a guess of course, I haven’t done the analysis. And Gary doesn’t show any such analysis in his book or white papers.

  2. Joe (arebelspy) November 15, 2015 at 7:03 pm #

    Good post. Appreciate the honest review (and longer timeframe), even if it’s not favorable.

    I think it’d be tough for me to stick with this portfolio type, even if I believed in it. Hope it works out for you.

    • Miles Dividend M.D. November 15, 2015 at 8:05 pm #

      Thanks Joe. I appreciate the comment.

      Incidentally if you want more honest reporting and diversity of thought on the MMM forum, you know where to find me.


      • chaostrader November 15, 2015 at 9:13 pm #

        Good reply.

        I was thinking about other “sell signals”, like only switch when GEM tells me to sell two months in a row, or skip the last month, which should still get one out of bear markets early enough. But the underperformance issue is part of the strategy and it will only change when the shit hits the fan again like it did in 2000 and 2008.

        • Miles Dividend M.D. November 15, 2015 at 9:45 pm #

          Yeah Chao.

          Negative tracking error is a really a fact of any active strategy. Just ask Warren Buffett circa 1999. You have to deal with false positive signals to get out, in order to harvest the benefits of true positive signals to get out. There is a lot to be said for keeping it simple, I think.


  3. Fred C Dobbs November 16, 2015 at 11:57 am #

    Gary does show in his absolute momentum paper (which is included in his book) that a 12 month look back gives higher Sharpe ratios.

    • Miles Dividend M.D. November 16, 2015 at 12:13 pm #

      But not for dual momentum.

      Change the assets, and test for yourself. There will always be a highest Sharpe ratio for one look back period for any group of assets with DM, but unless there is a significant difference in the Sharpe ratio for all or most asset pairs, with one lookback period significantly yielding higher Sharpe ratios, this result can be perfectly random and of no significance.

      If 12 months is more efficient, then you would expect to see a pattern where 12>11>10>9….in terms of sharpe ratio.

      When I looked at this informally, that was not the case. But admittedly this is anecdote. Please look at this yourself, and report back. It’s an interesting question.

  4. Lord Metroid November 19, 2015 at 7:17 pm #

    I started momentum investing, I simply sort in descendning order the funds that has had the best track record for the last year and choose to invest in the best. Very easy and satisfying. I also keep an eye out for commodities like Silver, Gold, Oil, etc.

    • Miles Dividend M.D. November 19, 2015 at 7:39 pm #

      If one of the funds is cash then we’re both DM investors!

  5. Ming November 20, 2015 at 7:38 am #

    Good to see an update on this. Interesting to note that it was better than how your previous portfolio would’ve done. And that seems like a positive.

    After reading the book and it does seem that the real gains would come from another large and deep bear market the likes of 2000/2008.

    ZeroHedge just published an article that hints at what may be the next crash due to the abnormal behavior of the markets in concert with the Fed’s QE. Namely that the QE prop up of markets has led to a distorted psychology within the market that is driven entirely by how much the Fed is propping up things rather than more market fundamentals: http://www.zerohedge.com/news/2015-11-19/fed-has-set-stage-stock-market-crash

    • Miles Dividend M.D. November 20, 2015 at 8:31 am #

      Ming, there is no question that big bear markets are where DM kills the market. But they are by no means the sole area of outperformance. In antonacci’s book GEM handily outperforms both the domestic and foreign stock markets from 1973 to 1999, and recall that the biggest bear in the 80’s was the flash crash of 87 which is the worst case scenario for DM, and during which DM got killed.

    • Miles Dividend M.D. November 20, 2015 at 8:37 am #

      As to the source of the next crash, the only thing I’m confident about is that I have no idea what will cause it. In addition I am very suspicious of those who claim to have the ability to predict the future behavior of markets. These predictions tend to tell me much more about the biases of the prognosticator than about the future itself.

  6. Joe November 21, 2015 at 8:54 pm #

    Thanks for the insightful post. I have been trolling this blog for a while after reading the Dual Momentum book and finding your site.

    I was pretty confident that the strategy would be sound given the behavioral reasons behind it (Thinking Fast and Slow was excellent), but I have to admit that the recent volatility has got me questioning if I could stomach finding myself on the wrong side of the trades over time…

    I definitely look forward to future posts on the topic! Thanks for all you do

    • Miles Dividend M.D. November 22, 2015 at 12:01 pm #

      Thanks Joe,

      There is no way around the fact that periods of negative tracking error will impact any strategy that is not invested in 100% home country stocks/bonds, including the 60/40 portfolio!

      This is obvious in the abstract, but the experience of losing to “the market” for months at a time is psychologically tougher than it might seem.

      More posts coming, stay tuned.


  7. Kyle November 22, 2015 at 10:39 am #

    It sounds like there is a lot of data mining going on here. Be careful. The problem is that there will always be issues. As soon as you think you have the ultimate look back period and it works once or twice you will think you are golden…guess what ? The markets change and evolve. Trend following I think will continue to work in the future, but will work differently. It may not be as profitable. There are many trend following methods that used to work when nobody did it like buying a 50 day breakout and selling a 20 day low. But backtests show that it doesn’t work well anymore. That was a Momentum strategy, but the markets evolved, got smarter and rejected it. Dual momentum may work but we may find that for the next 10 years a 18 week lookback was the way to go. The market is getting more volatile and you are going to get more whipsaws and sell when you think there is a big crash coming only to see a full recovery in a couple weeks. The market is mean reverting, especially volatile ones like the USA. In smaller international markets trend following works a little better when less volatility. I like dual momentum, but markets are changing fast and furious. Be prepared for years of whipsaws and losses. My point is, stop fiddling with lookbacks and data mining, who knows how the future will be, my guess is in the next 20 years a different lookback will prove to be more optimal. Hindsight is always 20/20. There is no risk in the past. We all have a very bumpy ride ahead. My advice is make dual momentum a small part of your portfolio, no more than 10-20%. Diversify your strategies. Thanks for sharing your results!

    • Fred C Dobbs November 22, 2015 at 11:08 am #

      Gary talks about the tendency to over optimize in his latest blog post:


      The 12-month look back he uses has held up well since it was introduced in 1937 without there needing to be any data mining. There will always be some whipsaws and occasional under performance with dual momentum, but I’ve never seen anything else as good on a risk-adjusted basis. People who don’t appreciate that and who ignore dual momentum or keep it as only 10-20% of their portfolio are a reason why it should continue to work well in the future. It will never be over exploited.

      • Miles Dividend M.D. November 22, 2015 at 11:56 am #

        He’s right of course. But the truth is that all trading periods between 3-12 months work fairly similarly, aside from the increased trading frequencies inevitably assocoated with shorter lookback periods, which I agree is significant. Which lookback will perform beat after trading costs for the next 20 years is anybodies guess, but I would not bet against 12 months.

  8. Grant November 22, 2015 at 11:21 am #

    A very interesting follow up. I haven’t gone live yet, but have been tracking how things would have worked out for the last year. I’ve been using a 12 month look back (because Gary recommended that), large cap ETF’s (as Gary also recommended) and living in Canada, having read an article for foreign investors, have been using US ETFs for stocks and Canadian Bonds. I have been tracking VV and ACWX against BIL. Interestingly,during the last 12 months, I have stayed in US equities. I had set my monthly look back day as the 1st of the month. If the look back day had been September 30th or Oct 2nd, I would have gone to bonds for a month, but October 1st kept me in US equities. Just one day made that difference!

    • Miles Dividend M.D. November 22, 2015 at 11:47 am #

      I imagine you would have killed your current buy and hold portfolio for the last year right?

      But yeah, trading decisions are often random depending on the day of trading. But this noise should wash out over time.

      One approach would be to create 1-2% bands around your trade decisions witha bias towards NOT trading. Ie if a different asset doesn’t out perform your asset by at least 1-2%: dont trade. My bias is to keep it simple though.

  9. Kyle November 22, 2015 at 11:26 am #

    “I’ve never seen anything else as good on a risk-adjusted basis”

    “It will never be over exploited”

    I guess the holy grail had been found! Nothing can go wrong! It can’t miss! It’s a sure thing! No red flags about those quotes above….

    Here is what will happen. This strategy is made to prevent huge bear markets so you can get out with a loss of 20% instead of 50%. What if the market doesn’t look like the last decade? What if all the market corrections are only 25% and no more. With this strategy you will sell at the worst time and lock in losses.

    Economists can only explain the mania of the Internet and tech stocks in the late 90s or the real estate bubble in 2008 as a time of irrational exuberance. Maybe that was the refresher and it won’t happen for another 25 years.

    I am not saying dual momentum is bad. I like it too. But to assume the market won’t evolve and reject ideas that have worked in the past and to assume this is a holy grail and it can never be over exploited is just naive and immature. The smart thing to do and what any risk manager would do is don’t rely on just one thing. Don’t base your whole future on one investment strategy. You can do whatever you want, just like people can continue to do drugs, drink, smoke and not exercise . It’s your life, you choose what to make of it. If you want to bet your life on dual momentum go for it. Those a little more worldly will know to manage your risk with more than one strategy.

    By the way…comments like the quotes above is what always gets investors in trouble. I’m sure real estate was a ” sure thing ” that will ” never be over exploited” in 2007, as was gold once, as was airline stocks. The list goes on.

    As soon as you think you are golden and smarter than the market, it will humble you.

    Be smart, don’t be naive.

    • Miles Dividend M.D. November 22, 2015 at 11:43 am #

      There are no sure things. That’s for sure.

      But the achilles heel for dual momentum will be the same as for any strategy, I believe; behavioral errors. Negative tracking error is a behavioral problem with any strategy, including buy and hold. A lot of people talk about their willingness to absorb the volatility of 100% stocks, but I’m dubious. Multiple studies show that people dramatically underperform the funds they invest in. I certainly wouldn’t trust myself with 100% stocks B and H.

      A quick point though. Although 200-2010 was the best decade for DM, this does not mean that the strategy is wholly dependant on big bear markets to succeed. One need only look at GEM’s performance from 1972-1999 to know that this is not the only secret sauce. Diminishing drawdowns is the most important factor, but responding to regime change with relative momentum is also value added.

  10. Mike April 12, 2016 at 12:41 pm #

    I’ve traditionally been a buy and hold the cheapest ETF you can find sort of guy, but it seems like you might be interested in what Wesley Gray and the folks over at Alpha Architect are doing with momentum. They have a couple of books out, I’ve read DIY Financial Advisor and plan to read Quantitative Momentum and Quantitative Value. Their Momentum strategy is more based around individual securities, which it has occurred to me you could execute reasonably cheaply in taxable accounts using Motif Investing, but they also have an actively managed Momentum ETF, and a Value ETF. You might be able to leverage these if you don’t want the hassle of managing Motifs or individual securities.

  11. Dennis May 12, 2016 at 8:25 am #

    Don’t want to be a salt in the wound kind of guy.

    However, isn’t the research pretty clear that if you are going to be a stock investor, then actively managed stuff doesn’t beat index fund investing?

    I don’t invest in the stock market, but when I did…..that is the information that I got over and over again during my due diligence.

  12. Joao Antunes September 23, 2017 at 11:33 am #

    Curious to know how is your DM portfolio doing 3 years later?

    • Miles Dividend M.D. September 23, 2017 at 2:14 pm #

      It’s doing great. Actually beating the S&P over past year which ain’t easy. I need to do an update. Stay tuned!

Leave a Reply

Visit Us On TwitterVisit Us On FacebookVisit Us On Google Plus