Share this Article & Support our Mission Alpha for Impact
This is some text inside of a div block.
This is some text inside of a div block.
This is some text inside of a div block.

Momentum: The most important findings.

Post by 
Text Link
Cross-Sectional Momentum has long been known on the stock market as a yield anomaly. Nevertheless, the vast majority of scientific studies show that the effect still exists. This article summarizes the most important findings from capital market research on this topic in chronological order.

Avoiding the value effect

The paper "Does the Stock Market Overreact?" published in 1985 [1] examines the price development of winner and loser stocks. From this, conclusions can be drawn for the momentum effect. The study shows that, over a comparatively long ranking period of 5 years, loser stocks perform significantly better on average than winner stocks. This means that momentum stocks that have been rising or falling at above-average rates for a very long time should be avoided. These stocks are increasingly drifting into the value range, which represents an increasing reverse risk for momentum stocks.

avoid value effect
Figure 1) Avoid value effect
The chart shows the course of the cumulative average returns for loser and winner portfolios following a 5-year ranking period. Overlapping portfolios were examined in the period from 1933 to 1978 (monthly data).
Source: DeBondt, W. / Thaler, R. (1985), Does the Stock Market Overreact?, Journal of Finance, Vol. 40, No. 3, p. 803

Use momentum life cycle

In 2000, the study "Price Momentum and Trading Volume" [2] described the concept of a momentum life cycle for the first time. This is a model of the presumed interaction of price momentum, reversal and trading volume. According to this model, winner stocks (loser stocks), which have a high (low) volume, are in the late phase of their momentum movements and are therefore more likely to experience a reversal than another pro-cyclical movement in the short term. Conversely, winner (loser) stocks with low (high) volume are in the early phase of their momentum cycle, so the procyclical movements are more likely to continue here. However, the authors point out that low-volume stocks generally perform better than high-volume stocks.

momentum life cycle
Figure 2) The momentum life cycle
Winner (loser) stocks with low (high) volume are in the early stages of their momentum cycle according to the model, so that procyclical movements are more likely to continue than reverse. However, the authors explicitly point out that this is a rough concept that only applies at the portfolio level.
Source: Lee, C. M. / Swaminathan, B. (2000), Price Momentum and Trading Volume, Journal of Finance, Vol. 55, No. 5, p. 2063

The paper "Trading Volume and Momentum: The International Evidence" [3] published in 2015 confirms the life cycle model and shows that trading volume is related to the strength and persistence of the momentum effect. The researchers investigate the early-momentum strategy, which is long in low-volume winner stocks and short in high-volume loser stocks. Conversely, the late momentum strategy is long in high volume winner stocks and short in low volume loser stocks.

The results show that the early-momentum strategy is more profitable than the classic and late-momentum strategies. It achieves 1.22% per month on average across all 37 countries studied, outperforming the other two strategies by 0.38% and 0.74% respectively. Furthermore, the researchers show that the persistence of the momentum effect can be predicted with the help of trade volume. Thus, the early-momentum strategy remains profitable for 5 years on average across all countries, while the late-momentum strategy shows a significant reversal after only one year.

use the momentum life cycle
Figure 3) Use momentum life cycle
The chart shows the cumulative monthly returns for early and late momentum and the classic momentum strategy.
Source: Bornholt, G. N. / Dou, P. / Malin, M. (2015), Trading Volume and Momentum: The International Evidence, Multinational Finance Journal, Vol. 19, No. 4

Observe consistency of return development

The study "Predicting Stock Price Movements from Past Returns: The Role of Consistency and Tax-Loss Selling" [4] published in 2004 already showed that Momentum Winning is about consistency. According to the study, expected returns are higher when a past momentum movement is caused by a series of consistently positive months rather than by very few, very good months. Possible causes for this observation are slow information diffusion and the disposition effect. According to the authors, uniform momentum movements also represent a proxy for low volatility, which as an independent anomaly also has a positive effect on returns.

The 2014 paper "Frog in the Pan: Continuous Information and Momentum" [5] also provided evidence that momentum movements characterized by a continuous flow of information continue and do not form a reversal. Thus, a multitude of small pieces of information is more lasting than rare, clear messages. On average, a momentum movement takes about 8 months after a continuous flow of information, whereas it takes only 2 months on average after clear news.

The authors discuss an underreaction as an explanation: This is caused by information below the perception threshold of market participants being priced in with a delay. An indication of such sustained momentum movements is provided by a high proportion of positive, but low to moderately high daily returns.

consistency of returns
Figure 4) Observe consistency of return development
The chart illustrates the typical difference in price development between continuous (much small information) and discrete information flow (little large information). Both stocks have the same start and end price, but behave very differently in the meantime.
Source: Da, Z. / Gurun, U. G. / Warachka, M. (2014), Frog in the Pan: Continuous Information and Momentum, Review of Financial Studies, Vol. 27, No. 7

Focus on media-presented shares

The study "Media Makes Momentum" [6] published in 2014 shows the astonishing effect that the media actually increase the momentum instead of reducing it as expected on the basis of the expected diffusion of information (more information = more efficiency). According to the study, the media primarily influence the winner and loser stocks on the market, but hardly influence the midfield of the Momentum ranking. Moreover, the influence is stronger for winner shares than for loser shares.

It is interesting to note that the effect is even stronger when winners (losers) whose media presence has a positive (negative) tone are specifically considered. One explanation for this is that momentum movements lead to more media presence in the first place. This presence then often triggers a new reaction to information already known, which is then corrected again in the long term.

stocks with media coverage
Figure 5) Focus on media shares
Shown are the momentum returns for stocks with high and low media coverage. First the top/flop 20% of the coverage was determined and then a top/flop 30% momentum strategy was applied.
Source: Hillert, A. / Jacobs, H. / Müller, S. (2014), Media Makes Momentum, Review of Financial Studies, Vol. 27, No. 12

Avoiding Momentum Crashes

The 2016 paper "Momentum Crashes" [7] shows that devastating drawdowns in momentum are partly predictable. They occur when a panic phase develops after strong market slumps, followed by a significant upward countermovement. According to the authors, a major cause of these crashes is the variable beta of portfolios. In strong downward movements, the winner stocks (in the long portfolio) have low and the loser stocks (in the short portfolio) high beta values. If the market then suddenly turns upward, the result is a clearly negative long-short beta overall, which leads to a collapse in returns. The main problem here is a so-called "junk rally" on the short side. The historically most extreme examples were July and August 1932 and March to May 2009, when the 12-month loser portfolios rose by 232% and 163% respectively, while the winner portfolios only rose by 32% and 8% respectively.

momentum crashes
Figure 6) Avoiding momentum crashes
The markings show the extremely strong short-term performance of the loser portfolio towards the end of the two major bear markets. This was the main cause of the respective momentum crashes.
Source: Daniel, K. / Moskowitz, T. J. (2016), Momentum Crashes, Journal of Financial Economics, Vol. 122, No. 2, p. 224

Exploiting strong and weak momentum phases

In 2017 the study "One Brief Shining Moment(um): Past Momentum Performance and Momentum Reversals" was published. 8] This study examines the connection between recent past momentum performance (Past Momentum Performance, or PMP for short) and the returns of unchanged momentum portfolios. It shows that momentum portfolios created in the top 20% PMP months show a strong reversal over a 2 to 5-year period. Conversely, Momentum portfolios created in Flop 20% PMP months show permanently positive returns.

A possible explanation for this is short-term performance chasing by market participants, which leads to above-average momentum returns and thus top PMP months. Ultimately, this results in an overreaction that is later reduced again. Conversely, low PMP values can be justified by low interest in momentum strategies, which is associated with an underreaction and later stable higher returns. Investors can therefore use particularly strong (weak) momentum phases countercyclically to reduce (build) new positions.

weak and strong momentum phases
Figure 7) Use strong and weak momentum phases
The figure shows the PMP effect. The higher the past momentum returns, the stronger the long-term reversal.
Source: Ali, U. / Daniel, K. / Hirshleifer, D. (2017), One Brief Shining Moment(um): Past Momentum Performance and Momentum Reversals, Columbia Business School Research Paper No. 17-48

Find stocks with accelerated earnings

The paper "Earnings Acceleration and Stock Returns" [9] published in 2018 shows that an acceleration of earnings from quarter to quarter significantly contributes to the explanation of future excess returns. It examines a strategy that is long (short) in the 10% of stocks with the highest (lowest) acceleration. In each case, the 2nd day after the announcement of the corresponding figures is decisive. Over the period from 1972 to 2015, this retrospective calculation achieves an average market-adjusted return of 1.8 for US equities over a one-month period and 3.4% over a three-month period. The largest part of the effect is thus generated within the first month.

Yields can be further improved. To this end, the authors define a strategy in which the long side shows a high earnings acceleration and positive earnings momentum, while the short side only trades those stocks that, in addition to a low earnings acceleration, previously showed positive earnings growth on an annual basis, which has now turned negative. The 1-month return of this strategy is 2.6%.

One possible explanation, according to the authors, is that the market does not (partially) price in predictable effects of earnings acceleration for the resulting earnings growth over two to three quarters.

stocks with earnings acceleration
Figure 8) Find stocks with earnings acceleration (1978-2015)
Shown are the average market-adjusted, market capitalization-weighted returns of the 10% highest (Decile 10) and lowest (Decile 1) earnings acceleration as well as the corresponding long-short returns for up to 30 days after the announcement of the earnings figures.
Source: He, S. / Narayanamoorthy, G. (2018), Earnings Acceleration and Stock Returns, Working Paper, Tulane University, p. 33

Optimize time of day for transactions

In 2019 the previous working paper "A Tug of War: Overnight Versus Intraday Expected Returns" was published in a journal. 10] The authors show the astonishing effect that momentum returns on average occur outside trading hours. This is a parallel finding to the yield phenomenon in the overall market, for which other studies have shown average positive overnight returns.

A plausible explanation for overnight momentum returns is the general tendency of institutional market participants to trade against momentum characteristics during trading hours, which creates a certain price pressure. This effect can be exploited by generally opening new positions at the closing price of a day and closing existing positions for opening.

momentum vs trading hours
Figure 9) Optimise time of day for transactions
The chart illustrates that the momentum premium is mainly incurred outside trading hours. A 12-1-x momentum strategy in the US market with holding periods of 1 to 60 months was examined.
Source: Lou, D. / Polk, C. / Skouras, S. (2019), A Tug of War: Overnight Versus Intraday Expected Returns, Journal of Financial Economics, Vol. 134, No. 1, pp. 192-213

Conclusions

The findings from momentum research are helpful as possible filter criteria and can be combined with each other in systematic trading strategies. If many criteria are applied, negative screening is of interest in order to exclude the respective share of stocks with clearly undesirable characteristics for each criterion and thus obtain a pool of pre-selected potential Momentum stocks. Furthermore, common sense criteria can be used - for example, to focus only on the long side and only if there is a superior bull market.

Overall, systematic trading strategies based on momentum should in practice focus on both the return and risk dimensions. Relevant previous studies have been cited for both dimensions. It is recommended that new research findings be continuously integrated into the process in order to maintain and ideally expand existing advantages over other market participants.

[1] DeBondt, W. / Thaler, R. (1985), Does the Stock Market Overreact?, Journal of Finance, Vol. 40, No. 3, pp. 793-805
[2] Lee, C. M. / Swaminathan, B. (2000), Price Momentum and Trading Volume, Journal of Finance, Vol. 55, No. 5, pp. 2017-2069
[3] Bornholt, G. N. / Dou, P. / Malin, M. (2015), Trading Volume and Momentum: The International Evidence, Multinational Finance Journal, Vol. 19, No. 4, pp. 267-313
[4] Grinblatt, M. / Moskowitz, T. J. (2004), Predicting Stock Price Movements from Past Returns: The Role of Consistency and Tax-Loss Selling, Journal of Financial Economics, Vol. 71, No. 3, pp. 541-579
[5] Da, Z. / Gurun, U. G. / Warachka, M. (2014), Frog in the Pan: Continuous Information and Momentum, Review of Financial Studies, Vol. 27, No. 7
[6] Hillert, A. / Jacobs, H. / Müller, p. (2014), Media Makes Momentum, Review of Financial Studies, Vol. 27, No. 12, pp. 3467-3501
[7] Daniel, K. / Moskowitz, T. J. (2016), Momentum Crashes, Journal of Financial Economics, Vol. 122, No. 2, pp. 221-247
[8] Ali, U. / Daniel, K. / Hirshleifer, D. (2017), One Brief Shining Moment(um): Past Momentum Performance and Momentum Reversals, Columbia Business School Research Paper No. 17-48
[9] He, S. / Narayanamoorthy, G. (2018), Earnings Acceleration and Stock Returns, Working Paper, Tulane University
[10] Lou, D. / Polk, C. / Skouras, p. (2019), A Tug of War: Overnight Versus Intraday Expected Returns, Journal of Financial Economics, Vol. 134, No. 1, pp. 192-213

Would you like to use this article - in full or in part - for your purposes? Then please consider the following Creative Commons-Lizenz.