Beat the Market (Part 2) - Understanding and Using Momentum Investing
Momentum in finance has been described as the "premier anomaly". In regular person speak, that means the PhDs (and Nobel Prize winners) haven't figured out why it still makes money despite the market being "efficient". It's an annoying reality that conflicts with elegant theory.
In its simplest form, momentum investing is a system that buys securities with high returns in the past, and sells those with poor returns. This is based on the phenomenon that winners (securities that performed well relative to peers) continued to outperform on average, and losers (securities that performed relatively poorly) tended to continue their underperformance.
While not as popular an investment style as value or growth, momentum has been used by practitioners for decades. One of legendary trader Jesse Livermore's rules, as detailed in Edwin Lefevre's 1923 classic Reminiscences of a Stock Operator, was to buy stocks that were rising and sell those that were falling.
Richard Driehaus of Driehaus Capital Management, a current-day practitioner of momentum investing, reportedly delivered 30% annual returns over 12 years after setting up his firm in the 1980s.
"I would much rather invest in a stock that’s increasing in price and take the risk that it may begin to decline than invest in a stock that’s already in a decline and try to guess when it will turn around." – Richard Driehaus, the "father" of modern momentum investing
The evidence for momentum
Academic research on momentum has picked up in the past three decades, and there are now many studies showing the efficacy of momentum investing.
In a seminal 1993 paper Jegadeesh and Titman showed that strategies which buy stocks that performed well in the past and sell stocks which performed poorly generated significant positive returns over the next 3 to 12 months. For example, over the period from 1965 to 1989, buying stocks based on their past 6-month returns and holding them for 6 months generated a compounded excess return of 12.01% per year above the market on average.
The excess return of momentum has been observed across a broad range of geographies, asset classes and time periods:
"The existence of momentum is a well-established empirical fact. The return premium is evident in 212 years of U.S. equity data (from 1801 to 2012) — as well as U.K. equity data dating back to the Victorian age in over 20 years of out-of-sample evidence from its original discovery, in 40 other countries and in more than a dozen other asset classes. Some of this evidence predates academic research in financial economics, suggesting that the momentum premium has been a part of markets for as long as there have been markets." - from Fact, Fiction and Momentum Investing (Asness, Frazzini, Israel and Moskowitz, 2014)
As far as backtests go, it's hard to beat this one showing that momentum works within and across different asset classes over the last 215 years: Two Centuries of Multi-Asset Momentum (Equities, Bonds, Currencies, Commodities, Sectors and Stocks) (Geczy and Samonov, 2017).
Momentum doesn't work just as a standalone factor – it can also be integrated with other lowly correlated strategies such as value to improve the risk-adjusted returns of a portfolio.
Despite all the evidence showing momentum has worked, it is a widely misunderstood and much-maligned phenomenon. A good paper looking at the popular arguments against momentum and the evidence refuting them is Fact, Fiction and Momentum Investing (Asness, Frazzini, Israel and Moskowitz, 2014).
Not all academic research has found evidence in support of momentum1. There are also some non-academic momentum skeptics as well, who attribute the phenomenon to factors such as randomness and rebalancing.
Why does momentum work? And will it continue to work?
In short, nobody knows for sure2.
Some academics think momentum investors face higher risks using this strategy, and therefore get higher returns as compensation. Others question whether momentum strategies actually outperform once transaction costs are taken into account3. Still other academics theorize that momentum investors are exploiting the behavioral mistakes of other investors.
Whether the explanation is risk-based or behavioral-based (or a mix of the two) is less important because if either (or both) were true, it suggests that the momentum premium will continue to persist4.
Some investors look to Newton's laws of motion as an explanation, applying the laws of classical mechanics to the markets. As an economics major I'm going to butcher it, but simply put Newton's first law states that an object in motion tends to stay in motion unless an external force compels it to change, i.e. the principle of inertia5. For momentum investors, this means that a stock going up will tend to keep going up, and vice versa, unless there is a change in the forces of demand and supply causing it to move up.
Issues with momentum investing
Investors should be aware of several issues when using momentum to invest.
First, turnover tends to be high in momentum strategies, so transaction costs could eat away at returns unless you trade highly liquid securities and with a broker that charges low commissions.
Second, it could be difficult for individual investors to benefit from momentum via funds. A 2018 report by Dimensional Fund Advisors showed that all but one of the public momentum funds had underperformed the Russell 3000 broad market benchmark from June 2003 through to 2017.
Third, momentum doesn't work in all market conditions and it can underperform for long stretches of time. A 2018 Morningstar report found that the momentum factor could underperform its benchmark by more than a decade.
Fourth, momentum doesn't work equally well across all types of securities. In a 2011 research paper Gary Antonacci looked at relative momentum applied to equities — single stocks, sectors, styles, and countries. He found the best results with global stock indices. This has been backed up by the Geczy and Samonov (2017) paper linked above as well.
Finally, a simple momentum strategy of buying winners and shorting the losers can come with large drawdowns, i.e. there is significant crash risk. High-flying momentum stocks tend to fall the hardest during big corrections. It is a volatile strategy that can be hard to stick to.
Mitigating drawdown risk with dual momentum
Traditional momentum as studied in academia typically looks at the returns that can be generated by buying the stock market's recent winners and shorting the recent losers. This strategy of looking at relative momentum, also known as relative strength, views price strength in comparison with other securities.
We can also compare the price of a security relative to its own trading history. Absolute momentum, also known as time-series momentum, looks at the past price performance of an asset to infer future movement.
Gary Antonacci, an investment professional with over 40 years' experience and an award-winning researcher, introduced the brilliant concept of dual momentum in a 2012 paper, using both relative momentum to get the excess return it brings, while combining it with absolute momentum to reduce volatility and drawdown, protecting investors from extreme losses during bear markets (his book, Dual Momentum Investing: An Innovative Strategy for Higher Returns with Lower Risk is highly recommended).
On his website Gary shows an almost 70 year (1950-2018) backtested result of his Global Equities Momentum (GEM) model, which holds U.S. or non-U.S. stock indices when stocks are strong, and uses bonds as a shelter when stocks are weak.
The results are impressive. Over this time period, his GEM model outperformed a Global Asset Allocation (GAA) benchmark comprising 45% U.S. stocks, 28% non-U.S. stocks, and 27% 5-year bonds (the amount of time GEM spent in each of these markets) by a large amount (5.8% per year), with a higher Sharpe ratio (0.96 vs. 0.57) and much lower drawdowns (worst drawdown of 17.8% vs. 41.2%).
In case you're wondering, GEM also beat the S&P 500 index over this period (by 4.4% per year), with a favorable Sharpe ratio (0.96 vs. 0.52) and worst drawdown (-17.8% and -51.0%).
In my own implementation of a systematic (i.e. rules-based) momentum strategy, I've relied heavily on the concept of dual momentum, and am grateful to Gary for his contribution to this field.
In the next few parts of this series I will look at the building blocks of my own strategy, and how you can use it to beat the market.
Footnotes
- A recent paper, Momentum? What Momentum? (Theissen and Yilanci, 2020), concludes that the momentum effect may be much weaker than previously thought.
- An academic paper that tries to classify all the various explanations for why momentum works is Equity market momentum: A synthesis of the literature and suggestions for future work (Subrahmanyam, 2018)
- When looking at a sample set of data for stocks, the likely outliers in terms of price performance (i.e. those that went up or down the most) are usually small, illiquid stocks with high transaction costs. Combined with the high turnover that momentum strategies usually have, that could significantly reduce (or even eliminate) the outperformance. But (as is the case for my "Beat the Market" strategy) this argument is not valid for broad market ETFs which are very liquid.
- If you're interested in this topic I'd recommend reading Explanations for the Momentum Premium (Moskowitz, 2010)
- We could probably stretch the analogy to cover the second law (i.e. Force = Mass x Acceleration is analogous to relative momentum) and third law (i.e. "For every action, there is an equal and opposite reaction" is analogous to the tug-of-war between momentum and mean reversion investors) but that'd probably be veering into pseudoscience.