Updated Dual Momentum Test

I frequently get asked for updated tests on various strategies. Using Portfolio123 I ran a backtest on a Dual Momentum strategy from 1/1/2007 – 5/25/2016. The strategy is updated on Scott’s Investments monthly, the most recent update is here.

The strategy invests equally in one ETF from each of four baskets of ETFs/cash:

  1. Equities – VTI, EFA, or Cash
  2. Credit Risk – CIU, HYG, or Cash
  3. Real Estate Risk – VNQ, REM, or Cash
  4. Economic Stress – GLD, TLT, or Cash

The backtest below ranks each ETF on 12 month momentum and purchased the ETF with the highest 12 month momentum.  The 12 month return of the ETF must also be positive in order for it to be purchased, otherwise cash is held.

Re-ranking every 4 weeks results in the following return statistics.  No reduction is made for commissions or taxes:




8 thoughts on “Updated Dual Momentum Test”

  1. With only one-quarter of the dual momentum assets allocated to equities, it doesn’t look like the S&P 500 is the best benchmark to use.

  2. It seems that, other than the lower drawdown, there isn’t much to gain from this?
    I notice more and more that techniques which initially show excellent performance revert to average, or worse, in real time investing.

    I’d appreciate your views, Scott.


  3. “Lower drawdown / volatility is the primary draw”

    Actually, Gary Antonacci states that from 1974 to 2015 his dual Momentum strategy returned 17.3% annually, vs 12.2% for the S&P.
    Also a vast reduction in drawdown, as you say.

    1. Stephen, it’s tough/impossible to go from paper model to actual real trading and get 17% with lower drawdown and a good sharpe. Also why should we assume the market will act similar over the coming years.

      I suspect anyone that could capture 17% for the next 40 years would become Buffetesque. Gary is a smart guy and I think Dual Momentum might be on to something that maybe can beat the market on a risk adjusted level.

      But if you really want/hope for 17% plus going forward you will have to find a cheap way to leverage these strategies. Heck,most people think even 7% over the next 10 years will not be too easy (read Hussman).

  4. Full year’s data for HYG and CIU is available only since 2008.

    How did you do the calculations starting in 2007?

  5. Just because something goes through a slow spell for a period of time (even 2-3 years) doesn’t mean it’s not valid any longer. I notice more and more that many investors have a short attention span and bail out on any strategy when it gives poor results for a year or two. This is why the average investor loses money, or gets much lower returns than the market over time. Nothing works all the time. The question you have to ask is when it doesn’t work well will it do me great harm? (Think: value investing drawdowns of 70% when “value” isn’t working.) If you can avoid the big losses (not all losses) and make reasonable money in other years, over the long run YOU WIN! My only adaptation to this system would be to ditch the 12 month number, and consider trading quarterly.

Comments are closed.