AgPa #37: Momentum Investing – Fact and Fiction

Fact, Fiction, and Momentum Investing (2014)
Clifford Asness, Andrea Frazzini, Ronen Israel, Tobias Moskowitz
The Journal of Portfolio Management Special 40th Anniversary Issue 2014, 40(5) 75-92, URL/AQR

After examining the general Facts and Fictions about factor investing last time, this week’s AGNOSTIC Paper examines the first factor in more detail. I follow the authors’ order of publishing and start with momentum.

Everything that follows is only my summary of the original paper. So unless indicated otherwise, all tables and charts belong to the authors of the paper and I am just quoting them. The authors deserve full credit for creating this material, so please always cite the original source.

Setup and Idea

What is momentum? I have given a quite comprehensive overview about it in my first Seeking Alpha article on a momentum ETF, but let me summarize it again. To set the stage, it is important that the authors (and thus this article) focus on cross-sectional momentum. This is important because time-series momentum (a.k.a. trend following or managed futures) is a completely different type of strategy.

The idea of cross-sectional momentum is surprisingly simple. According to the authors, it is “[…] the phenomenon that securities that have performed well relative to peers (winners) on average continue to outperform, and securities that have performed relatively poorly (losers) tend to continue to underperform.” For the most basic form of this strategy, you just rank securities by their 12-month returns and go long (short) a portfolio of the securities with the highest (lowest) returns. Historically, such a strategy produced meaningful risk-adjusted returns across many different samples and is a robust empirical fact.

Despite the compelling evidence, there are some interesting myths around momentum and the authors attempt to kill them in this overview-paper. One brief disclaimer before we start. The article is from 2014. That means it lacks the insights of the last 8 years of research on momentum. In my opinion, however, the authors picked fairly timeless issues which are still relevant (and valid!) today. The fact that those ideas survived another 8 years is by itself a nice result. To enter the nerdy-speak of quantitative investing, the explanations and ideas of this paper worked out-of-sample…

Data and Methodology

Similar to last week, there is not so much to say about specific data or methodology. The paper is an overview about momentum investing and some myths around it. That said, the authors use a lot of data and references to back up their arguments. To be as transparent as possible, they use the well-known and publicly available momentum factor from Kenneth French’s website for most of their analyses.

Important Results and Takeaways

Myth 1: Momentum returns are economically not meaningful

According to the authors, some people claim that momentum returns are too small and sporadic to rely on. Given that the evidence for momentum is so compelling, this is a quite strong statement. The authors quote a lot of research showing that momentum worked over more than 200 years, in both international and US stock markets, and even within other asset classes. They also show that momentum returns were historically more reliable than those of the aggregate stock market. So if you believe momentum is sporadic, you must also stay away from the even more sporadic equity risk premium (which most investors obviously don’t do!).1The authors examine the fraction of rolling 5-year-returns between 1927 and 2013 to make this argument. The stock market delivered positive returns 82% of the time, the momentum factor stands at 88%.

The authors conclude (and I fully agree) that this myth simply ignores the data. In fact, momentum is probably one of the most pervasive factors of all and even Eugene Fama, an outspoken skeptic with respect to potential inefficiencies, acknowledges that momentum is “[…] the premier anomaly.”2Quote from Fama & French (2008).

So momentum is not sporadic. But is it economically meaningful? Over the long-term (here 1927 to 2013), the momentum factor produced an annualized return of 8.3%.3Remember that the momentum factor is a hypothetical long-short portfolio, not a long-only equity strategy. Over the same time, the realized equity risk premium was 7.7%. So in terms of raw returns, momentum was considerably better (60 basis points over 80+ yeas compounds nicely!). The same is true for risk-adjusted returns. The Sharpe ratio of the momentum factor stands at 0.50 versus 0.41 for the aggregate stock market. Further note that with these numbers, momentum was historically also stronger than the value and size factor. The picture obviously looks different for other sub-sample periods. But overall, raw and risk-adjusted returns in those spheres are definitely meaningful.

Myth 2: Long-only investors cannot capture momentum

Most academics and hedge funds look at momentum from a long-short perspective. According to the authors, some people therefore question whether momentum can also be captured by long-only investors. Short answer: it can. Going long and short is conceptually nothing different than over and underweighting positions with respect to a long-only benchmark. In the first case, you hope to earn an absolute return from the spread between momentum-winners and losers. For the second, you hope to earn an excess returns with respect to a benchmark. In fact, I have written a series of Seeking Alpha articles about long-only ETFs that successfully implemented momentum for long-only investors.

The authors also show, again using Kenneth French’s momentum factor, that the long and short side of the momentum factor contributed roughly the same to the overall performance in the period from 1926 to 2013. They also cite research showing similar patterns for a 40 year sample of international equity returns and different asset classes. The bottom line remains the same: long-only investors can also capture momentum.

Myth 3: Momentum is much stronger among small-caps

This myth is actually quite thoughtful. It is well-researched that most factor strategies work better with small caps and there are several plausible explanations for this. For example, prices of small caps could be less efficient as large sophisticated players cannot employ too much capital on them. In addition to that, there is simply a larger number of small stocks out there. Quantitative strategies generally work best with large universes where you can place as many bets as possible. From this perspective, small caps offer a much larger opportunity set.

Although this myth applies to most other factors, it is less extreme for momentum. The authors split Kenneth French’s momentum factor by market capitalization and show that momentum returns were historically quite similar across small and large caps. It still worked a little better for small caps, but the difference is not large enough to conclude that momentum is much stronger among among small caps.

Myth 4: Momentum does not survive trading costs

Once again, this myth is not completely wrong. Momentum is one of the “faster” factors and comes with higher turnover than other strategies. Therefore, it is completely reasonable to ask whether appealing academic backtests can be actually turned into profitable real-world strategies. In fact, this is a question investors should ask themselves about any strategy, not just momentum. If you find an ultra-profitable strategy that requires intra-day trading, chances are high that it won’t survive real-world transaction costs.

So what about momentum? Based on more than a trillion dollar of live trading data from AQR Capital Management (the authors’ employer and one of the most prominent momentum investors in the world), the authors estimate real-world trading costs of momentum strategies.4They actually reference a research paper that examines real-world trading costs of the most common factor strategies. The result? Implementation costs of momentum strategies are lower than previously thought. There are also several techniques to mitigate transaction costs even further and the evidence suggests that asset managers use them. Overall, the results therefore clearly suggest that momentum remains profitable after costs. Another, more simple way, to arrive at this conclusion is to look at real-world track records of momentum funds. And there are enough out there which delivered momentum profits after implementation costs and management fees.

Myth 5: Momentum produces a huge tax bill

This one is closely related to the high-turnover problem from the previous myth. All else equal, higher turnover increases the probability of a higher tax bill because you realize more profits on the way. However, the authors bring a whole line of arguments why this is not a problem for momentum strategies.

First, by construction, momentum sells losers early and holds winners for longer. From a tax-saving perspective, this is pretty efficient and should actually lower the tax burden of momentum. Second, empirical analyses show that momentum tends to have lower exposure to dividend-paying stocks which is again tax-efficient. Third and finally, the authors cite other research showing that you can easily consider tax issues in an optimized momentum strategy without losing the overall character of the strategy. Bottom line: momentum not only survives trading costs but also tax payments. If anything, it is actually quite efficient with respect to taxes.

Myth 6: Momentum is better as a screen than as a factor

In my opinion, this is the strangest of the ten myths. The authors argue that some people don’t regard momentum as a stand-alone factor but would be still willing to use it as a screen on top of other strategies. The authors rightfully question the logic behind this statement. If you don’t believe that momentum (a.k.a. past performance) predicts future stock returns, why would you use it as a screen for other strategies?

The only plausible way to defend this is that people somewhat belief in momentum, but not strong enough to use it without other data. In what is probably the best quote of the paper, the authors argue that such a behavior is similar to being “a little pregnant”. If you believe in momentum you should be willing to put capital behind it and if you don’t, you shouldn’t be willing to use it as a screen.5You can obviously scale your personal momentum exposure according to your individual beliefs. That is fine and not the issue this myth is aiming for.

Myth 7: Momentum returns should disappear in the future

The standard concern with all strategies that are nowadays well-known. Why should factors continue to work in the future if we all know about them? This was also one of the key points in last week’s post. Momentum, as virtually all other factor strategies, is not arbitrage. Even though attractive over the long-term, momentum strategies live through painful crashes and long drawdowns. So it is not always easy to stick with it. As I mentioned last time, asking why factors continue to work despite being well-known is similar to asking why people in the rich-world still suffer from obesity although most of us know about unhealthy food. The optimal behavior is pretty obvious, but it is not easy to actually do it.

In the paper, the authors bring an even more fascinating argument. As we know from last week, there is compelling evidence that combining multiple factors (value, momentum, defensive, carry) is even more appealing than each one individually. In particular, momentum was historically negatively correlated with value. That means a good year for momentum tends to be a bad one for value and vice versa. A combination of the two is therefore even more attractive because the combined portfolio is much more robust.6You combine two things that do well individually but at different times. A combination of the two should therefore do reasonably well all of the time (in theory!).

Having this in mind, the authors run a portfolio optimization for a multi-factor portfolio involving value and momentum. The surprising result: even if the momentum premium disappears (0%) or turns slightly negative (down till -3% per year), momentum would still be attractive within a multi-factor portfolio because of its diversification benefits. In fact, all of the arguments to combat the previous myths are even stronger when evaluating momentum within a multi-factor context.

Bottom line: both empirical and logical evidence does not indicate a disappearing momentum premium. With the benefit of hindsight in 2023, it is fair to say that this prediction was accurate and momentum continued to work over the last years. And as the little exercise on portfolio optimization shows: even if the myth would be true (it is still not!) and the momentum premium indeed disappears, rational multi-factor investor should be still willing to use the strategy for its diversification benefits.

Myth 8: Momentum is too volatile to rely on

The authors argue that this myth is most likely a result of recency-bias.7The psychological effect to overweight more recent experiences and forgetting about the true distribution of outcomes. Momentum suffered a painful crash in 2009 during the recovery after the financial crisis. As mentioned in the last myth, this shouldn’t come as a surprise. Momentum is not a true arbitrage strategy and to collect the premium over the long-term you have to occasionally live through such painful periods.

The authors still argue that it is unfair to discard momentum by cherry-picking one particular painful period. The aggregate stock market also suffered from painful crashes in the past and many people are still willing to invest in market-tracking index funds. In addition to that, they again explain that when combined with value in a multi-factor portfolio, the momentum crash of 2009 wasn’t as painful. Bottom line: just because a strategy has some bad periods doesn’t mean it is too volatile altogether. Look at such periods to prepare yourself for what may be coming, but don’t overweight them when evaluating strategies.

Myth 9: Different momentum measures lead to different results

Similar to some other myths, this one is not entirely false. Different momentum measures will obviously lead to different results. However, the authors explain that most people use this myth to argue that momentum is not a robust strategy and doesn’t survive deviations in methodology.

While it is always required and beneficial to stress-test strategies with different ways of measuring the same phenomenon, the authors show that this is unlikely a problem for momentum. The simplest measure of all, the past 12-month return without the most recent month, worked in many different out-of-sample studies thus indicating that momentum is indeed a pervasive empirical phenomenon. On top of that, many smart researchers came up with different momentum definitions that obviously doesn’t lead to the exact same results, but to a very similar pattern – past winners outperform past losers on average.

Myth 10: There is no theory behind momentum

This final myth is closely related to the question why momentum should continue to work even though we all know about it. As I mentioned before, the key argument are plausible reasons why momentum profits exist in the first place. In this final part, the authors provide some examples.

Explanations and theories behind factor strategies generally fall into two categories: rational/risk-based or behavioral. Academics and practitioners still debate which of the two prevails, but investors have the luxury that they don’t need to play this game. For us, it is sufficient that there actually are some plausible theories. In reality, most of them are probably active at the same time anyway.

Behavioralists argue that momentum profits come from imperfect rationality of some investors. In English: some investors are for whatever reason willing to take the other side of the momentum trade and thereby give up returns over the long-term. The most prominent examples for such explanations are the initial underreaction (investors only gradually incorporate new information into the price) and the delayed overreaction (investors chase returns and bid-up the price of winner securities after the fact).

Risk-based explanations, in contrast, argue that momentum profits are a fair compensation for some kind of systematic risk. This could for example arise from economic patterns of the underlying businesses. Another possibility are the aforementioned momentum crashes and drawdowns. No one wants to underperform, but especially not during the recovery after a crisis. Bearing the risk of underperforming when everyone else recovers is therefore a plausible explanation for a long-term premium.

The most important point remains, and this is brilliant to conclude this post, that there actually are plausible theories why momentum exists. Together with the compelling empirical evidence, I think we can be reasonably sure that momentum is not just statistical noise. Another reason for this is the fact that real-world quantitative investors like AQR Capital Management successfully implemented the strategy for now almost 30 years out-of-sample.

This content is for educational and informational purposes only and no substitute for professional or financial advice. The use of any information on this website is solely on your own risk and I do not take responsibility or liability for any damages that may occur. The views expressed on this website are solely my own and do not necessarily reflect the views of any organisation I am associated with. Income- or benefit-generating links are marked with a star (*). All content that is not my intellectual property is marked as such. If you own the intellectual property displayed on this website and do not agree with my use of it, please send me an e-mail and I will remedy the situation immediately. Please also read the Disclaimer.

Endnotes

Endnotes
1 The authors examine the fraction of rolling 5-year-returns between 1927 and 2013 to make this argument. The stock market delivered positive returns 82% of the time, the momentum factor stands at 88%.
2 Quote from Fama & French (2008).
3 Remember that the momentum factor is a hypothetical long-short portfolio, not a long-only equity strategy.
4 They actually reference a research paper that examines real-world trading costs of the most common factor strategies.
5 You can obviously scale your personal momentum exposure according to your individual beliefs. That is fine and not the issue this myth is aiming for.
6 You combine two things that do well individually but at different times. A combination of the two should therefore do reasonably well all of the time (in theory!).
7 The psychological effect to overweight more recent experiences and forgetting about the true distribution of outcomes.