Frequently Asked Questions
It seems that a lot can happen between the beginning of the month when your momentum indices adjust their positions and the beginning of the next month when new signals are given.
Research has shown that stock prices are reactive rather than trending for shorter periods of time, such as days and weeks. Stocks are more likely to rebound from short-term fluctuations rather than continue in the same direction. It is usually better to ignore short term movements and retain a longer term perspective.
If momentum investing is so great, why aren’t more people doing it?
It usually takes awhile for academic research to work its way into the investment marketplace. We saw that with indexing and value investing. Public awareness should grow as information about momentum gets better assimilated over time.
There are also other reasons why momentum investing has not caught on. First, it is not as psychologically appealing to buy things that have already gone up in price as it is to buy things that appear to be cheap. Second, momentum may seem too simple to some people who are think more complex approaches have more merit. Finally, some investors have strong prejudices toward other styles of investing.
Interestingly, the same behavioral factors that explain why momentum works can also explain why momentum has not caught on yet: anchoring, conservatism, and the slow diffusion of information, along with herding by those who are are still strongly attached to other investing approaches.
What the main risks associated with dual momentum investing?
One risk is that past history is driven by randomness, and the future may not be like the past. However, there are now several hundred years of out-of-sample performance supporting momentum and behavioral reasons why the momentum effect has a good chance to continue to work.
There is some risk that momentum will attract substantial capital and not work as well in the future because of too many participants. But indices are much more scalable than individual stocks.
Another risk are the occasional timing miscues that cause dual momentum portfolio to lag behind its benchmarks. Over the past 40 years, our Global Equities Momentum (GEM) model under performed its benchmark in 1979-80 and 2009-11. No strategy outperforms all the time.
There is also risk associated with the trend-following component of dual momentum being slow moving in order to minimize whipsaws.This means that much of the short term volatility of the stock market still exists with dual momentum.
There may be re-entry lags when a new bull market begins after dual momentum has taken you out of a prior bear market. Please see the Disclaimer page of this website for additional risk factors.
Should one invest using your papers’ momentum portfolios?
My papers were meant to illustrate the principles of momentum investing and not as models for actual investing. Historically, the highest risk premium by far has come from equities. Research shows that momentum works best with geographically diversified stock indices.
My book’s GEM model is designed as a do-it-yourself approach using dual momentum focused on stock indices. The section called “How to Use It” in Chapter 8 of my book shows how you can easily implement the GEM model.
How many sectors does your Dual Momentum Sector Rotation (DMSR) model invest in at any one time?
The number of sectors can vary if you use a buffer zone for exits. The number of sectors can also depend on one’s attitude toward reward and risk. Fewer sectors mean higher expected returns but higher volatility. You can use the principles in my book to develop your own sector rotation model that best suits your own risk preferences. However, on a blog post I show that the performance of DMSR using additional data since my book was published is no longer as attractive as the performance of broader-based dual momentum models like GEM. I no longer use sector rotation nor encourage others to use it.
Why do you update the performance of only your Environmental, Social, and Governance and Global Equities Momentum models on your website now?
My ESGM and GEM models are publicly disclosed. Anyone who reads my book can easily implement them. My other dual momentum models have continued to do well but are proprietary. I license them to several investment professionals.
I like your Global Balanced Momentum (GBM) model because it holds both stocks and bonds. What can I do along those same lines?
My GBM model is different now from what was published in my book. I license it and my other proprietary models to investment professionals. To achieve a similar effect, you can use the suggestion at the end of chapter 8 of my book and allocate a permanent percentage of your portfolio to aggregate bonds and the remainder to GEM.
What ETFs do you use in your Global Equities Momentum model?
ETFs for the S&P 500 index are IVV or VOO. (SPY is a unit investment trust that has restrictions on it, such as no security lending or reinvesting of dividends. It therefore has a slight return disadvantage.) ETFs for non-U.S. stocks are VEU or VXUS. The ETF for U.S. Treasury bills is BIL (or you can use SHV), and ETFs for aggregate bonds are BND or AGG.
Are there any outside validations of your GEM model?
Have you looked at whether there is a best time of the month to reevaluate and rebalance your portfolios?
There are minor performance differences based on what day you use. But these may just be normal data noise. Rebalancing around month-end has given good past results. It is important to stay with whatever day you decide to use and not be swayed by emotion to change your signals.
During the past year, GEM has had some close signals for switching in and out of stocks. Is there some rule of what to do when this happens?
All our published results are based strictly on month-end closing prices. I have tested using a 50/50 allocation to stocks and bonds when the signal is close until the following month. This does not make much difference in long run risk-adjusted returns. However, to minimize feelings of regret no matter what happens, you may want to use this approach.
It looks like the GEM model has under performed the S&P 500 since 2009. Has momentum lost its effectiveness?
Selection bias can lead one to choose starting points for performance evaluation that can give results that are either too positive or too negative depending on one’s inclination. This apparent under performance is actually attributable to using an inappropriate starting date. You should not separate the performance of 2009 from the performance of 2008. Here is why.
Dual momentum has a trend following component that lags behind when a new bull market begins.That is the cost of avoiding the carnage of the preceding bear market. It is unfair to look only at the new bull market without also considering the preceding bear market.You need to look at a full market cycle at least.
From 2008 through 2014 the S&P 500 index was up 63%, while GEM was up 84%. An added benefit is that GEM avoided the stress of substantial bear market losses in 2008.
Has the FED policy keeping interest rates low for so long had any effect on dual momentum profitability?
Quantitative easing and low interest rates have helped boost stock returns. Whenever there are extreme returns in stocks, it is difficult for any trend following approach that exits the market occasionally to keep up with it. Low interest rates have also meant low returns when dual momentum was in bonds.
Stock valuations look high now, and bond yields are low. Will this adversely affect dual momentum returns?
High stock valuation levels can mean lower expected stock returns, and low bond yields point to lower future bond returns. But stocks and bonds can still fluctuate and create opportunities. In 2000, there were also high stock valuations and low yields. But our GEM model had a compound annual return of 11.6% over the next 10 years, while a 60/40 stock/bond portfolio returned 2.3%
Chart courtesy of SharpeReturns.ca
Will dual momentum will lose its effectiveness when more people start using it?
An anomaly can certainly lose profitability if it starts being widely followed. However, the behavioral basis behind momentum has been strong and persistent. Human nature does not readily change, and so I expect the momentum effect to continue to persist.
It is unlikely that most people will become dual momentum believers. Many have been brought up as value or buy-and-hold investors. Their biases may keep them from adopting dual momentum.This is especially true of institutional investors who often have a strong aversion to any kind of tactical allocation. Even if this were not the case, Shleifer and Vishny (1997) show that asset managers are afraid of strategies that deviate from their benchmarks because investors may leave them following periods of underperformance.
In aggregate the performance of actively managed funds is inferior to passively managed funds. People have known about that for many years. Yet over 70% of all domestic equity funds are still actively managed. Momentum investing may very well show the same disconnect.
How do you determine the best look back periods for your models? Won’t shorter look back periods get you into and out of the markets sooner?
Academic research shows that over the long run, momentum for stocks works well with a look back period of 6 to 12 months. I like to use longer look back periods to minimize transaction costs, increase the likelihood of long-term capital gains, and reduce the number of false signals.
While shorter look backs may get you out of and back into markets sooner, they can produce more whipsaw losses that actually lead to lower returns and larger drawdowns. A 12-month look back period was found to work well by Cowles & Jones in their 1937 momentum study. It has held up well ever since and is commonly used in other academic research. Staying with this look back reduces concerns about data mining.
What about combining different look back periods?
Both volatility and return are reduced some when you do this. Overall, I have found no advantage over using a single look back period. Keeping things simple is often the best policy.
Some investors skip the last month when applying momentum. Why don’t you do that?
It makes sense to skip the last month when you are applying momentum to stocks because individual stocks can overreact to news and mean revert. If you want to use momentum with stocks (not recommended), see the book Quantitative Momentum. If you use stock indices or other asset classes, you do not need to skip the last month.
Why does the performance of the GEM model begin in 1974 in your book and 1971 on your website?
When my book was written, the earliest bond index data I could get was from 1973. Now I have additional bond index data, so the GEM performance record on my website goes back to January 1971 where it is now constrained by the lack of index data for international stocks. I can go back to 1927 when looking at U.S. equities with absolute momentum where international stock index data is not needed. See my blog posts, Absolute Momentum Revisited and And the Winner Is… In addition, both relative strength and absolute momentum have been individually tested back to 1801. See Geczy and Samonov (2015).
When I back tested GEM, there were a few months in which I see GEM is in foreign stocks, but you show GEM in aggregate bonds. Why is that?
On page 101 and 112 of my book I give a simplified logic for GEM so anyone can easily implement it using a free charting website. There is a minor difference if you calculate the signals as discussed on page 98 of my book. There I mention how I determine absolute momentum using only the S&P 500 index, since the U.S. leads world equity markets. I cite a supporting reference. This means we may occasionally be in aggregate bonds if the trend in U.S. stocks is down even when non- U.S. stocks are the strongest asset. I have always used the later method for generating GEM signals.
Isn’t value investing the opposite of momentum investing? How can both be valid?
They operate on different time frames. Value is based on long-term mean reversion, while momentum relies on intermediate-term (usually 3 to 12 months) serial correlation. Despite all the attention it has received, there is strong evidence that value investing has not held up well well in the real world. See my factor investing blog post.
Why do you use absolute momentum for trend following instead of moving averages?
With moving averages you compare an asset’s current price to an average of its prices over the look back period. With absolute momentum you compare the current price to the price at the beginning of the look back period. This means moving averages signals are more sensitive and give more than 25% more trades than absolute momentum. Absolute momentum thus has fewer false signals and whipsaw losses, as well as lower transaction costs.
Zakamulin shows that absolute momentum outperforms 5 different types of moving averages on 145 years of out-of-sample data. It is one of only two methods that outperforms the market with statistical significance.
What do you think of using stop-losses with dual momentum?
Stop-losses were once thought to reduce returns when they reduce risk exposure. Recent research shows that stops can actually enhance returns if they are used with care. I have a blog post called “Momentum and Stop Losses” that discusses this. Dual momentum is more effective than stop losses in reducing risk exposure while also enhancing returns. Stops are redundant and unnecessary when using dual momentum.
Since momentum works best geographically, why not use it with country index ETFs?
Dual momentum works best when volatility is not too high. Individual countries can have very high volatilities. This can make it difficult to get in and out of them using trend following momentum without giving up much of their profit. This is the same reason we do not use small cap indices or stock sectors.
How do leveraged ETFs perform with dual momentum?
There are drawbacks to using leveraged ETFs. First, most leveraged ETFs use daily resets which are best suited for day trading. There can be large tracking errors when holding leveraged ETFs on a longer term basis. Daily resets are also not tax efficient since leveraged ETFs will result in mostly short term capital gains or loses. Finally, leveraged ETFs have a leverage factor of 2X or greater. This may be too much leverage. There is still considerable short term volatility with dual momentum that may cause discomfort with 2X or greater leverage. A more attractive alternative to leveraged ETFs might be for aggressive investors to search for a brokerage firm offering low margin borrowing rates.
Have you looked at using inverse equity instead of bond ETFs when absolute momentum
tells you to exit stocks?
Equities are the core of our models because they have shown the highest long-run risk premium. Shorting stocks is therefore climbing an uphill battle. We want every advantage we can get by having risk premium on our side to serve as a tailwind for future performance.
There are also higher costs associated with inverse ETFs. You can own an S&P 500 ETF for an annual expense of 4 basis points, while the expense ratio of an S&P 500 inverse ETF can be 89 basis points. Also, because stocks have an upside bias and our models are slow moving, there is often not much profit from short positions by the time you enter and exit. The average bull market duration since 1942 has been 32 months, while the average bear market has lasted only 12 months. Switching to bonds during stock market weakness as identified by dual momentum has historically done better than being short stock indices.
There are international stock ETFs that hedge their currency exposure. What do you think about using these with GEM?
There is a tendency for International stocks to outperform U.S. stocks when the U.S. dollar is weak and non-U.S.currencies are strong. This lets us profit from the strength in non-U.S. currencies. When non-U.S. currencies are weak, GEM is often out of international stocks. So there is little reason to use hedged ETFs. GEM automatically deals with exchange rate risk.
Chart courtesy of SharpeReturns.ca
What is the best way for non-U.S. Based investors to use dual momentum?
Gogi Gerwal has a guest post on my blog called “Dual Momentum for Non-U.S. Investors” that discusses this in detail. The only idea I would add is that non-U.S. Investors may want to include their home country in their portfolio.
Do you make your buy and sell decisions using month-end prices?
Consistent with most academic work, my models’ entries and exits are based on monthly closing prices that reflect total returns. If you want to trade the same day as your signals instead of waiting until the next day’s open, you can set up a portfolio on SharpCharts that updates real time during trading hours.
What are momentum crashes, and do I need to worry about them?
These are caused by the short side of long/short momentum suffering large losses when stocks rebound sharply off severe bear market bottoms. This happened in 1932 and 2009. Since we do not hold short positions, momentum crashes are irrelevant to us.
How do you invest your own funds?
I developed dual momentum specifically for my personal investing. Nearly all my liquid net worth is invested using the same dual momentum models that I license to investment advisors.
With bonds being in a bull market over the past 35 years, does the use of aggregate bonds with Global Equities Momentum (GEM) overstate future expected performance?
Aggregate bonds have an average duration of only around 5 years. This means they are not as sensitive to interest rate changes as longer duration bonds. As the following chart shows, their returns have been relatively stable and steady over the past 40 years. According to Nuveen, income, rather than price appreciation, made up 93% of the total return from aggregate bonds.
Furthermore, GEM and ESGM have been out of equities only about 30% of the time, and aggregate bonds have been responsible for only 20% of these models’ profits. Using U.S. Treasury bills instead would have created 10% of our profits. So the contribution of bonds to our models’ returns was minor.
Absolute momentum is used to exit equities and enter bonds when it first identifies bear markets in stocks. These usually precede recessions, which often lead to falling rather than rising interest rates as aggregate demand lessens and the Federal Reserve tries to stimulate the economy. When stocks are weak, there is also a tendency for investors to move from stocks to bonds, thereby increasing demand for bonds. So it seems prudent to accept a little duration risk then. If you disagree with this, then you can substitute shorter term bonds or Treasury bills for aggregate bonds without much fall off in performance.
© 2016 Portfolio Management Consultants All Rights Reserved