The authors note empirical studies have found that momentum
strategies have been found to exist across 2 centuries, several asset classes,
and geographies. However, there have
also been several rebuttals claiming that momentum strategies may not work effectively
for various reasons. The authors in this
paper intend to defend 10 myths about momentum strategies.
The first myth says that momentum strategies offer returns
that are too small and sporadic. The authors
reference several prior studies that show the robustness of momentum strategies
across several countries, regimes, and asset classes. They then use Ken French’s Up-minus-Down
(Winners minus losers) strategy over the 1927 – 2013 period, 1965 – 2013 period,
and the 1991 – 2013 period. Across all periods,
they find that this momentum strategy earned about 8%, significantly outperforming
other common factors such as RMRF (market premium), SMB (size premium), and HML
(value premium) strategies on both nominal returns and Sharpe ratios. The authors also examine the percentage of
annual periods and five year periods that the momentum strategy was
profitable. They find that about 80% of
the periods were profitable, while the other factors’ positive return percentages
were less. Finally, the authors built a
value/momentum portfolio, which outperformed the other factors even further on
both a Sharpe ratio and percent positive metrics. Therefore, the authors find this to be good
evidence to refute the claim that momentum strategies are small and sporadic.
The second myth says that momentum can only be exploited on
the short side (and is not very useful to long-only investors). The authors use the same time periods as in
Myth 1 and split the alphas of the returns of the UMB strategy into the long
and short pieces. In doing so, they find
that about half of the return relates to the profit from going long the past
winners, and half of the return is due to going short the past losers. The authors also reference a paper that finds
the same results over 86 years of US equity data, 40 years of international
equity data, and 40 years of data from other asset classes. Therefore, the authors find this to be good
evidence to refute the claim that momentum strategies are only profitable on
the short side.
The third myth says that momentum is only present in small-cap stocks and not in large-cap stocks. However, several researchers have found that small-cap and large-cap stocks contribute rather equally to momentum returns. The authors use Ken French's data to analyze the momentum returns (UMB) of large-cap versus small-cap stocks; in doing so, they find that small companies tend to have a larger momentum return than large-cap companies, but the difference is fairly small. Doing the same strategy using value/growth (HML), they find that small-cap companies have significantly higher returns than large-cap returns; and in fact, large-cap value premiums are significantly zero. So, it seems the momentum strategies do not prefer small over large-cap stocks as much as value strategies do. And actually, they note that if Fama/French had not normalized for company size, there would likely have been no value premium found at all. The authors also make note that Fama and French found in 2012 that over a 1989-2011 period in international equities, small cap momentum stocks had slightly higher returns than large cap momentum stocks, but the large-cap momentum returns were significant. This was also found in the 1927 - 2013 period used in the current paper.
The fourth myth says that momentum returns do not exceed trading costs. The thought is that momentum strategies have higher turnover and therefore higher trading costs. The authors use real-world trading costs from AQR Capital's trades over the 1998-2013 period in 19 developed equity markets across different factor strategies. They find that per-dollar trading costs for momentum strategies are actually quite low compared to other strategies (e.g., value, size, etc.). They also note that trading strategies should be used to reduce trading costs (e.g., limit orders rather than market orders). They find that prior studies using estimated trading costs often inflate them due to averages that include trades of retail investors (rather than just institutional investors). Using the AQR data, they find that since momentum strategy returns significantly exceed the returns of small-cap strategies and since trading costs of momentum strategies are lower than trading costs of small-cap strategies, the net returns to momentum strategies significantly exceed those of small-cap strategies.
The fifth myth says momentum strategies do not work for taxable investors; this of course relates to the higher turnover of momentum strategies. Prior research has found, however, that the tax burden of momentum strategies are about the same as value strategies, despite their having 5-6 times the turnover. This is due to momentum strategies instructing the investor to hold winners longer (which typically results in long-term capital gains status) and to sell losers (which results in deductions). Also, value strategies typically have high dividend exposure, which is taxed at ordinary rates; however, momentum strategies do not have as high of dividend exposure. As such, since momentum strategies' returns are significantly higher than value strategies' and they typically both have about the same tax burden, then momentum strategies outperform on an after-tax basis. Another item to note is that tax optimization is a lot easier to implement when concentrating on capital gains (because you can control when to sell securities) rather than dividends (because you can't control when you receive dividends).
The sixth myth says momentum strategies are best used for a screen rather than as an actual factor. However, it seems counter intuitive to say that momentum strategies are good or useful but to denounce it at the same time. This may be due to naysayers anchoring to the concept of market efficiency. The authors suggest that if all the previously discussed myths were true, it might be useful as a screen; but since they were debunked, we might say that momentum has more of a right to be an actual factor than value or size do.
The seventh myth says that investors should be worried about momentum returns disappearing. The authors argue that this should be the case for any factor, not just momentum. They argue, however, that momentum has had a more stable record than other factors over time (as was found in the first few myths in this paper). They note that naysayers may believe this myth due to the relatively newness of momentum studies in academia, and the use of behavioral reasons rather than risk-based causes of momentum. The authors remind us, however, that momentum has been found to exist going back 200 years and across dozens of equity markets; also, they note that any factor could be due to behavioral or risk-based factors (e.g., the promotion of value stocks could increase demand/price and reduce their premium to zero). The authors reference a 2013 paper that finds the out-of-sample period did not result in reductions to momentum returns (which might be evidence of continuity of the strategy). And importantly, the authors used Ken French's data to form portfolios of stocks with both momentum and value characteristics; they find that even if the returns of the momentum factor are zero, including them in the portfolio increases the sharpe ratio of the portfolio due to the diversification benefits.
The eight myth says that momentum is too volatile to rely on. However, we remember that the sharpe ratios of momentum strategies found in the previous myths are significantly higher than those of other factor strategies; and since volatility is taken into account when calculating the sharpe ratio, we know this myth to be untrue. The authors note the myth may come about due to the recent 2009 crash in momentum returns when the market significantly increased after the recession; in that instance, the momentum strategy would have held low-beta winners and high-beta losers, meaning the market would have moved up faster than the winners in the momentum strategy did. The authors note, however, that 1999 was bad for value investors and 2008 was bad for passive investors; so, they argue that an entire strategy should not be denounced just because of a few bad periods. They also note 1932 was bad for momentum and 1930 was bad for value. But, it was found that using momentum and value together during these bad times would have significantly reduced the volatility; using Ken French's data, the authors find that the worst drawdown for value-only was 43%, the worst drawdown for momentum-only was 77%, but when using value and momentum together, the worst drawdown was 30%. Also, being long-only would have also fared ok, because most of the loss in the momentum strategy is due to being short the high-beta stocks that suddenly increase in value.
The ninth myth says different measures of momentum can give different results over different periods. This is actually true, but may be a good characteristic, rather than bad. The argument of the naysayers is that data mining could be used to find the best strategy; but the authors note that out-of-sample evidence is robustly in favor of momentum strategies. Also, other factor strategies use different measurements as well (e.g., P/B, P/E, D/P, for value strategies). In the same vein, such metrics as past returns, past earnings, and analyst revisions are used in ranking momentum stocks; it was found in other studies, however, that each of these measures are effective independently and together. As such, this should be taken as a sign of robustness, rather than a critique, of momentum strategies.
The tenth myth is that there is no theory behind momentum. The authors note this is not fair, because other factors (e.g., value and size) do not have definite supporting theories either, and are in fact still heavily debated. The two behavioral reasons for momentum are that investors typically under-react to new information or delay-overreact to new information; both of these have been found to occur. The other possibility is that momentum occurs due to compensation for risk; for example, growth companies have a risk that they will not have enough cash to support their growth. The authors argue that no matter the reason that momentum occurs, there certainly does seem to be a persistence of momentum over time and should continue to exist going forward. In addition, the naysayers might be anchoring to the efficient market hypothesis and claim that the momentum premium should be arbitraged away; however, we've seen this to not be the case in the prior myths.
In conclusion, in this paper, the authors debunked 10 myths about momentum strategies; they welcome further debate around this paper or the effectiveness of momentum strategies.
About Sawyer Investment Management Company:
SIMCO is a Texas-registered Investment Adviser with its principal place of business in Dallas, Texas. It was formed on January 1, 2015 and is wholly owned by Ryan Sawyer, who is a CFA Charterholder and a Certified Public Accountant.
SIMCO specializes in the construction of equity portfolios, and is therefore an ideal resource for long-term investors. The firm goes through a rigorous process for selecting each and every holding in the portfolio. Rooted in the empirical research of academia, the portfolios are generally characterized as large-cap value momentum. For more information about how the portfolios are managed, see our website.
www.sawyerinvestment.com
https://www.facebook.com/Sawyer-Investment-Management-Company-1588110057913467/
https://twitter.com/SawyerInvest
The fourth myth says that momentum returns do not exceed trading costs. The thought is that momentum strategies have higher turnover and therefore higher trading costs. The authors use real-world trading costs from AQR Capital's trades over the 1998-2013 period in 19 developed equity markets across different factor strategies. They find that per-dollar trading costs for momentum strategies are actually quite low compared to other strategies (e.g., value, size, etc.). They also note that trading strategies should be used to reduce trading costs (e.g., limit orders rather than market orders). They find that prior studies using estimated trading costs often inflate them due to averages that include trades of retail investors (rather than just institutional investors). Using the AQR data, they find that since momentum strategy returns significantly exceed the returns of small-cap strategies and since trading costs of momentum strategies are lower than trading costs of small-cap strategies, the net returns to momentum strategies significantly exceed those of small-cap strategies.
The fifth myth says momentum strategies do not work for taxable investors; this of course relates to the higher turnover of momentum strategies. Prior research has found, however, that the tax burden of momentum strategies are about the same as value strategies, despite their having 5-6 times the turnover. This is due to momentum strategies instructing the investor to hold winners longer (which typically results in long-term capital gains status) and to sell losers (which results in deductions). Also, value strategies typically have high dividend exposure, which is taxed at ordinary rates; however, momentum strategies do not have as high of dividend exposure. As such, since momentum strategies' returns are significantly higher than value strategies' and they typically both have about the same tax burden, then momentum strategies outperform on an after-tax basis. Another item to note is that tax optimization is a lot easier to implement when concentrating on capital gains (because you can control when to sell securities) rather than dividends (because you can't control when you receive dividends).
The sixth myth says momentum strategies are best used for a screen rather than as an actual factor. However, it seems counter intuitive to say that momentum strategies are good or useful but to denounce it at the same time. This may be due to naysayers anchoring to the concept of market efficiency. The authors suggest that if all the previously discussed myths were true, it might be useful as a screen; but since they were debunked, we might say that momentum has more of a right to be an actual factor than value or size do.
The seventh myth says that investors should be worried about momentum returns disappearing. The authors argue that this should be the case for any factor, not just momentum. They argue, however, that momentum has had a more stable record than other factors over time (as was found in the first few myths in this paper). They note that naysayers may believe this myth due to the relatively newness of momentum studies in academia, and the use of behavioral reasons rather than risk-based causes of momentum. The authors remind us, however, that momentum has been found to exist going back 200 years and across dozens of equity markets; also, they note that any factor could be due to behavioral or risk-based factors (e.g., the promotion of value stocks could increase demand/price and reduce their premium to zero). The authors reference a 2013 paper that finds the out-of-sample period did not result in reductions to momentum returns (which might be evidence of continuity of the strategy). And importantly, the authors used Ken French's data to form portfolios of stocks with both momentum and value characteristics; they find that even if the returns of the momentum factor are zero, including them in the portfolio increases the sharpe ratio of the portfolio due to the diversification benefits.
The eight myth says that momentum is too volatile to rely on. However, we remember that the sharpe ratios of momentum strategies found in the previous myths are significantly higher than those of other factor strategies; and since volatility is taken into account when calculating the sharpe ratio, we know this myth to be untrue. The authors note the myth may come about due to the recent 2009 crash in momentum returns when the market significantly increased after the recession; in that instance, the momentum strategy would have held low-beta winners and high-beta losers, meaning the market would have moved up faster than the winners in the momentum strategy did. The authors note, however, that 1999 was bad for value investors and 2008 was bad for passive investors; so, they argue that an entire strategy should not be denounced just because of a few bad periods. They also note 1932 was bad for momentum and 1930 was bad for value. But, it was found that using momentum and value together during these bad times would have significantly reduced the volatility; using Ken French's data, the authors find that the worst drawdown for value-only was 43%, the worst drawdown for momentum-only was 77%, but when using value and momentum together, the worst drawdown was 30%. Also, being long-only would have also fared ok, because most of the loss in the momentum strategy is due to being short the high-beta stocks that suddenly increase in value.
The ninth myth says different measures of momentum can give different results over different periods. This is actually true, but may be a good characteristic, rather than bad. The argument of the naysayers is that data mining could be used to find the best strategy; but the authors note that out-of-sample evidence is robustly in favor of momentum strategies. Also, other factor strategies use different measurements as well (e.g., P/B, P/E, D/P, for value strategies). In the same vein, such metrics as past returns, past earnings, and analyst revisions are used in ranking momentum stocks; it was found in other studies, however, that each of these measures are effective independently and together. As such, this should be taken as a sign of robustness, rather than a critique, of momentum strategies.
The tenth myth is that there is no theory behind momentum. The authors note this is not fair, because other factors (e.g., value and size) do not have definite supporting theories either, and are in fact still heavily debated. The two behavioral reasons for momentum are that investors typically under-react to new information or delay-overreact to new information; both of these have been found to occur. The other possibility is that momentum occurs due to compensation for risk; for example, growth companies have a risk that they will not have enough cash to support their growth. The authors argue that no matter the reason that momentum occurs, there certainly does seem to be a persistence of momentum over time and should continue to exist going forward. In addition, the naysayers might be anchoring to the efficient market hypothesis and claim that the momentum premium should be arbitraged away; however, we've seen this to not be the case in the prior myths.
In conclusion, in this paper, the authors debunked 10 myths about momentum strategies; they welcome further debate around this paper or the effectiveness of momentum strategies.
Asness, Cliff S. and Frazzini, Andrea and Israel, Ronen and
Moskowitz, Tobias J., Fact, Fiction and Momentum Investing (May 9,
2014). Journal of Portfolio Management, Fall 2014 (40th Anniversary
Issue); Fama-Miller Working Paper. Available at SSRN: https://ssrn.com/abstract=2435323 or http://dx.doi.org/10.2139/ssrn.2435323
About Sawyer Investment Management Company:
SIMCO is a Texas-registered Investment Adviser with its principal place of business in Dallas, Texas. It was formed on January 1, 2015 and is wholly owned by Ryan Sawyer, who is a CFA Charterholder and a Certified Public Accountant.
SIMCO specializes in the construction of equity portfolios, and is therefore an ideal resource for long-term investors. The firm goes through a rigorous process for selecting each and every holding in the portfolio. Rooted in the empirical research of academia, the portfolios are generally characterized as large-cap value momentum. For more information about how the portfolios are managed, see our website.
www.sawyerinvestment.com
https://www.facebook.com/Sawyer-Investment-Management-Company-1588110057913467/
https://twitter.com/SawyerInvest
