Why value investing crushes momentum speculation

Overview

This article investigates fundamental assumptions – and refutes the key claim – of so-called momentum investors (who, as we’ll see, are actually speculators). Since the Global Financial Crisis, and particularly over the past several years, momentum approaches have become more prominent. Today, many “quant” funds and some ETFs utilise algorithms which purportedly detect and exploit momentum.

Momentum speculators buy stocks whose prices have recently been zooming and sell those which have been plunging. Value investors, in contrast, acquire stocks whose prices are low relative to fundamentals such as net assets, cash flows, dividends or earnings. Value investors seek to buy low and sell high; momentum speculators claim that they can generate superior returns by buying high and selling at even higher prices. 

Value investors buy and hold, and are thus low-turnover investors; momentum speculators, like most speculators, heavily “churn” their portfolios.

In The Intelligent Investor, Benjamin Graham counselled: “never buy a stock (just) because it has gone up or sell one because it has gone down.” His most famous and successful student, Warren Buffett, has been even more emphatic: “the dumbest reason in the world to buy a stock is because it’s going up” (see “Buffett Takes Stock,” The New York Times Magazine, 1 April 1990).

The core claims and practices of value investing and momentum speculation are thus incompatible.

What I’ve dubbed momentum speculation’s “strong form” contends that it handily outperforms value investing. Its “semi-strong” form overlooks their incongruity and asserts that momentum can be deployed advantageously in combination with value. “Fact, Fiction and Momentum Investing” by Clifford S. Asness, Andrea Frazzini, Ronen Israel and Tobias J. Moskowitz (The Journal of Portfolio Management, Fall 2014), presents a robust defence – indeed, a strident advocacy – of momentum speculation. I regard it as the clearest example of its “semi-strong” variant.

In this article, I analyse the same data (which Kenneth French compiled) that Ansess et al. did. I don’t, however, replicate their analysis: instead, I address its glaring – and, I believe, fatal – omissions. Asness et al. don’t, for example,

1. ascertain whether momentum actually exists, i.e., attempt to differentiate it from random fluctuation;
2. specify momentum’s average short-term (rolling 12-month) and medium-term (60-month) returns from year to year, under different market conditions, etc.;
3. compare the short-term and medium-term results of momentum speculation and buy-and-hold investing.

My analysis demonstrates that the key claim of “strong form” momentum speculation – that it outperforms “buy and hold” – is clearly false. Momentum virtually always underperforms; furthermore, the longer is the interval of time the more markedly it lags; and over the past few years it’s generated outright losses.

Why do momentum speculators consistently – and eventually massively – under-achieve “buy and hold” value investors? As I also show, the average duration of “momentum” is too short to distinguish from random fluctuation; moreover, its occurrence is too infrequent to generate outperformance – and, in many instances, decent results. Above all, by frequently “churning” their portfolios, momentum speculators forego the huge benefits, which increase the longer an investor holds her stocks, of compounding.

Most fundamentally, the assumptions of momentum speculation place unrealistic demands upon its practitioners. In particular, they can’t dependably distinguish “momentum” from random fluctuation. The “trend,” in other words, isn’t your friend: it’s a figment of your imagination.

Hence speculators can’t reliably select short-term winners – and because they churn their portfolios far more than buy and hold investors do, momentum speculators’ short-term winners can’t become long-term winners.

In short, momentum speculators usually don’t – because in principle they can’t – outperform. They’re speculators; speculators almost always underperform investors; hence buy and hold value investors usually outperform momentum speculators (see, for example, How Warren Buffett has trounced “the world’s greatest hedge fund manager,” 10 August 2025).

What Is “Momentum”?

Let’s start by clarifying a crucial reality: what’s commonly called “momentum investing” is actually “momentum speculation.” Investors strive to benefit from companies’ long-term operations; speculators, in contrast, seek to exploit short-term fluctuations of shares’ prices. 

Momentum, say Asness et al., “is the phenomenon that securities which 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.”

Their use of the word “relative” is significant. Momentum is similar to trend following (hence I’ll use “momentum” and “trend” interchangeably), but they’re not identical. “The process behind momentum,” Asness et al. elaborate, “is to rank securities relative to their peers; in contrast, trend following typically focuses on absolute price changes. Unlike trends, which increase exposure during upswings and decrease exposure during downswings, momentum takes no explicit view on the market trend, but simply ranks securities relative to each other over the same time period … Momentum’s ‘winners’ and ‘losers’ are defined no matter how the market overall is doing.”

Albeit in a relative rather than an absolute sense, momentum speculators’ attempt to profit from recent fluctuations of stocks’ prices – irrespective of companies’ fundamentals. Hence they’re clearly traders and speculators rather than investors.

Two actions distinguish them: they (1) avoid or sell assets (typically stocks, but in principle bonds, commodities, etc.) whose market prices have decreased comparatively sharply within the past 12 months, and (2) retain or buy assets whose prices have recently risen relatively strongly. Essentially, momentum speculators are betting that recent “momentum” (as I’ll demonstrate, it’s mostly indistinguishable from random fluctuation; behavioral economists dub it “herd mentality”) will persist.

Momentum speculators aim to “buy high and sell higher” by riding what they regard as short-term booms and dodging short-term busts.

Core principles of momentum speculation include:

• “The trend is your friend:” the core claim is that the trajectory of a stock’s price within the past year reliably predicts its course over the next 12 months. Specifically, so-called “winners” will continue to win and “losers” will keep losing.
• Exploitation of market inefficiencies: momentum speculators seek to exploit market inefficiencies and behavioral biases such as herd mentality and “fear of missing out” (FOMO). These emotions purportedly cause short-term price momentum to develop and persist; savvy speculators can allegedly profit thereby.
• “Technical analysis:” momentum speculation is a component of “technical analysis.” Although Asness et al. don’t, momentum speculators often rely upon price charts, volume and associated indicators. Not all momentum speculators are “technicians,” but most chartists are trend-followers. 

Finally, even its advocates acknowledge that momentum speculation doesn’t lack weaknesses. These drawbacks – which my analysis will ignore – include the need for (and for some people, increased anxiety resulting from) constant monitoring and heavy portfolio turnover; they also include larger – compared to “buy and hold” investing – transaction costs such as taxes on short-term capital gains.

In short, momentum speculation is potentially highly profitable; in practice, however, and like all speculation, it’s inherently risky and requires considerable discipline, active management, high tolerance of risk – and above all substantial luck.

Does Momentum Work?

Richard Driehaus (1942-2021) advocated the construction of portfolios comprising stocks whose prices have recently risen strongly. In 2000, Barron’s named him to its “All-Century” team of the 25 most influential people within the mutual fund industry over the past 100 years. He’s often been cited as the “father of momentum.” He famously asserted that “far more money is made buying high and selling at even higher prices.” Using this strategy, Driehaus Capital Management reportedly – I’ve not been able to locate data which corroborate this claim – delivered compound returns of 30% per year in the 12 years after its establishment in 1982.

Cathie Wood is the founder, CEO and chief investment officer of Ark Invest. She’s best known for her focus on “disruptive innovation;” by seeking to identify strong growth trajectories and market trends, her approach epitomises momentum speculation. Her flagship ARK Innovation ETF received extensive attention – and considerable adulation – from the mainstream media following its outperformance in 2017, 2020 and 2023.

On the other hand, Morningstar (“15 Funds That Have Destroyed the Most Wealth Over the Past Decade,” 2 February 2024) identified it as the industry’s third-highest “wealth destroyer” from 2014 to 2023.

Wood’s experience is hardly unique. Eric Ghysels et al. (“Momentum Trading, Return Chasing, and Predictable Crashes,” CEPR Discussion Paper No. DP10234, 10 November 2014) found that the increased risk of major losses accompanies momentum speculation. For example, in 2009 these portfolios crashed almost 75% within three months – substantially more than the overall market (see also Pedro Barroso and Pedro Santa-Clara, “Momentum has its moments,” Journal of Financial Economics, vol. 116, no. 1, April 2015).

Narasimhan Jegadeesh and Sheridan Titman conducted the first rigorous analysis of momentum speculation. They reported that buying stocks which have recently been outperforming and selling those which have been underperforming generates average returns of ca. 1% per month over the following 3-12 months (see “Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency,” The Journal of Finance, vol. 48, no. 1, 1993).

Others go much further; indeed, they’re unequivocal. “The existence of momentum,” assert Asness et al., “is a well-established empirical fact.” Its “return premium” exists across “multiple assets over 212 years of stock market data” in the U.S. They also assert that it’s existed since the mid-19^th century in Britain and currently in more than 30 countries. Accordingly, they allege that “the momentum premium has been a part of markets since their very existence, well before researchers studied them as a science.”

Meb Faber (“Relative Strength Strategies for Investing,” 6 April 2010) advocates the “strong form” of momentum speculation. He, too, is adamant: it allegedly “trounces a simple ‘buy and hold’ stock market strategy going back almost 100 hundred years.”

Data

“Momentum,” writes Burton Malkiel in A Random Walk Down Wall Street (W.W. Norton, 6^th ed., 2025), “is typically measured by looking at (a stock’s) last twelve months’ return excluding the most recent month. (The most recent month is eliminated because it often exhibits a reversal.) The measurement of the momentum factor is the average return of the best-performing 30% of stocks minus the average return of worst-performing 30% of stocks …” A momentum strategy thus entails “being long the best-performing stocks and short the poorest performers … No account is taken of trading costs, taxes and other possible implementation costs.”

In the words of Asness et al., “UMD, or ‘up minus down,’ represents a portfolio that is long stocks that have high relative past one-year returns and short stocks that have low relative past one-year returns.” UMD, in other words, captures momentum by going long short-term “winners” and short short-term “losers.”

A leading finance academic, Kenneth French, compiled the UMD data which Asness et al. and I analysed; these data quantify the momentum of American stocks since January 1927. Each month, French has admitted to his “momentum” portfolios the AMEX, NASDAQ and NYSE stocks which qualify and dropped those which cease to qualify. He records these portfolios’ total (that is, including dividends) return. I’ll refer to this average as the “momentum” portfolio.

Results

I’ve analysed the same data which underpin Asness et al.’s results but have conducted very different analyses. I’ve modified these data in one significant way: I’ve calculated returns on a CPI-adjusted basis, that is, removed the effects of consumer price inflation.

Very Long-Term

On a CPI-adjusted, total return (that is, and including dividends) basis, and ignoring transaction costs and taxes, each $1 invested in the momentum portfolio in January 1927 would have grown to $17.09 in October 2025 (Figure 1). In contrast, and on the same basis, each $1 invested in a portfolio which perfectly mimicked the S&P 500 Index would have grown to $957.

Figure 1: Investments of $1 in “Momentum” Portfolios and the S&P 500 Index, CPI-Adjusted, Total Cumulative Returns, January 1927-October 2025

The Index’s total proceeds are more than 56 times larger than the momentum portfolio’s The momentum portfolio’s compound annual growth rate (CAGR) is 2.9%; the Index’s is 7.1%. Over the entire interval from January 1927 to October 2025, the momentum portfolio has massively underperformed the Index.

Asness et al. are grossly mistaken: momentum generates no "long-term return premium." Quite the contrary: it produces a hefty deficit. Faber, too, is diametrically wrong: over almost a century a simple “buy and hold” stock strategy has utterly crushed “momentum.”

Short-Term and Medium-Term

Of course, virtually no family holds stocks or follows the same strategy for a century. Plenty, however, do over much shorter intervals. What about rolling short-term (12-month) and medium-term (60-month) intervals? What about the Index’s corresponding returns? For each month since January 1928 I’ve calculated the momentum portfolio’s and the Index’s CPI-adjusted, total 12-month returns, and for each month since January 1932 I’ve computed momentum’s and the Index’s CPI-adjusted, total 60-month returns (the latter expressed as CAGRs). Figure 2 and Figure 3 plot the results.

Figure 2: 12-Month Total Returns, CPI-adjusted, Momentum Portfolio and S&P 500, January 1928-October 2025

Over rolling 12-month periods, momentum’s return has averaged 4.2% and the Index’s 8.9%. Since 1928, momentum’s short-term underperformance has thus averaged 4.2% - 8.9% = -4.7% per year. In 716 of these 1,173 (61.0%) rolling 12-month periods, the “momentum” portfolio has underperformed. Both series have fluctuated essentially randomly around their respective means.

Figure 3: Five-Year, CPI-adjusted Total Returns (CAGRs), Momentum Portfolio and S&P 500, January 1932-October 2025

Over rolling 60-month periods (Figure 3), the momentum portfolio has averaged 2.7% per year and the Index 7.0%. Since 1932 momentum’s short-term underperformance has thus averaged 2.7% - 7.0% = -4.3% per year. In 786 (70.0%) of these 1,125 rolling 60-month periods, momentum underperformed the Index. The Index’s CAGRs have fluctuated approximately randomly around their mean; from the 1930s to the late-1970s, the momentum portfolio’s returns trended mildly upwards; since then, and particularly since the turn of the century, they’ve trended downwards.

Figure 4: 12-Month, CPI-adjusted Relative Total Returns, January 1928-October 2025

To quantify the momentum portfolio’s performance relative to the Index, for each month I’ve subtracted the Index’s 12-month and 60-month return from the momentum portfolio’s corresponding return; months when the relative return is below 0% indicate that the momentum portfolio has underperformed; months when the relative return is greater than 0% indicate that the momentum portfolio has outperformed. Figure 4 and Figure 5 plot the results.

Figure 5: 60-Month, CPI-adjusted Relative Total Returns, January 1932-October 2025

The 12-month relative returns (Figure 4) fluctuate randomly around their mean. The 60-month relative returns (Figure 5) are slightly parabolic: during the 1960s and 1970s, and again just before the Dot Com Bust and Global Financial Crisis, momentum outperformed; however, over the past 10-15 years it’s underperformed ever more markedly.

Does “Momentum” Actually Exist? Evidence from Runs Tests

Asness et al., Faber and momentum speculators assume that momentum exists; if it doesn’t, their contentions collapse. How to distinguish genuine momentum from mere random fluctuation? For each of the series (namely the rolling 12-month and 60-month average returns of the momentum portfolio and S&P 500 Index) I’ve conducted Wald–Wolfowitz runs tests. Named after statisticians Abraham Wald and Jacob Wolfowitz, this test ascertains the randomness or otherwise of a temporal sequence of data.

Among the ingredients of this test: a temporal sequence’s expected (under the assumption that runs reflect mere random fluctuation) number of runs, and its actual number.

A “run” is a segment of the sequence consisting of adjacent equal elements. Consider rolling 12-month returns since January 1928. If the return in the 12 months to February 1928 is greater than the return for the 12 months to January 1928, I code the former return “+” and if it’s less, I code it “-”. The result is an 1,173-item sequence of plusses and minuses. As an example, here’s a hypothetical 21-element sequence:

+ + + + − − − + + + − + + + + + + − − − −

It comprises six “runs.” Their lengths, respectively, are 4, 3, 3, 1, 6, and 4 elements. I’ve tested the null hypothesis that each element in the sequence has been drawn independently from the same probability distribution – in other words, the sequence contains no instances of actual momentum, but merely random variation.

Calculating the runs test generates expected and observed numbers of runs as well as other statistics; Table 1 summarises them. Two results are most significant: firstly, for both portfolios and over both short-term and medium-term intervals, the observed number of runs is less than – and for the momentum portfolio, is statistically insignificantly different from – the number we’d expect from random fluctuation; secondly, the length of the average run is very short.

Table 1: Characteristics of “Runs,” Average Momentum Portfolio and S&P 500 Index, January 1928-October 2025

The implication is fatal to momentum speculators: the occurrence of actual momentum – as opposed to random fluctuation – is too infrequent, and its duration too short, to generate outperformance (or even decent results).

To corroborate and elaborate this result, I’ve divided each series of data into two segments: positive runs (that is, consecutive occurrences of increasing rolling returns) and negative runs (consecutive occurrences of decreasing rolling returns). I’ve then stratified each segment by the run’s length: one month, two months, three months and four or more months. Finally, for each month of each run I’ve calculated the series’ mean 12-month and 60-month returns. Table 2 and Table 3 summarise the results (“N” means the number of observations in the sample, which dwindles as runs’ length increases).

For both the momentum portfolio and the Index, increasingly long positive runs generate ever higher short-term and medium-term returns (Table 2). However, examining the corresponding months of the momentum portfolio’s and the Index’s runs, without exception the momentum portfolio’s results underperform the Index’s.

Table 2: Returns from Positive Runs by Month, 1928-2025

This result suggests that, to the extent that momentum actually exists, it occurs more in the Index than in the “momentum” portfolio!

Conversely, for both the momentum portfolio and S&P 500 Index, increasingly long negative runs generate ever lower short-term and medium-term returns (Table 3). However, in almost all (19 of 20) instances the momentum portfolio’s results continue to underperform the Index’s.

Table 3: Returns from Negative Runs by Month, 1928-2025

Are Price Plunges a Signal to Sell – or Buy?

So far, I’ve established two crucial results: firstly, in absolute terms momentum speculation generates mediocre results; secondly, it usually – and cumulatively massively – underperforms the S&P 500 Index. Generally over the past century, “buying high and selling at even higher prices” hasn’t worked. Is that because “buy low and sell high” has?

For each month since January 1928, I’ve (1) calculated the S&P 500’s total, CPI-adjusted return during the previous 12 months, as well as its results (expressed as CAGRs) during the previous five years and next 12, 24, 36, 48 and 60 months; (2) rank-ordered the data by the return over the previous 12 months; (3) divided the dataset into five equal (by numbers of observation segments (“quintiles”); and (4) calculated the Index’s average CAGR over the next 12, 24, 36, 48 and 60 months. Table 4 summarises the results (returns which exceed their category’s average are green; cells below them are red).

Table 4: S&P 500’s CPI-adjusted Total Returns (CAGRs), by Quintile of Short-Term Past Returns, 1928-2025

Three results are paramount. Firstly, and unsurprisingly, the higher is the return over the past 12 months (second column), the higher has been the return over the past five years (third column). Secondly, the lower the return has been over the past year and five years, the higher, on average, are the subsequent returns: the future returns in Quintile #1 are generally higher than those in Quintile #2, those in Quintile #2 are higher than those in Quintile #3, etc.

Thirdly, this regression to the mean erodes but persists over time: a plunge of the Index at one point in time tends to be associated with superior returns not just one but also two, three, four and five years later; conversely, if the Index’s 12-month return zooms (Quintile #5) not only does it subsequently regress: five years later the return is relatively pedestrian.

Table 4 confirms that “buy low” works: the best time to buy the Index – from the point of view of subsequent above-average short-term and medium-term returns – is when it’s recently plunged. That’s the opposite of the modus operandi of momentum speculation.

I’ve replicated Table 4 with French’s momentum data. Table 5 summarises the results (again, the green cells exceed the categories’ means; the red ones are below them). As in Table 4, so too in Table 5: the higher the return over the past 12 months, the higher the return has been over the past five years.

Table 5: Momentum Portfolio’s CPI-adjusted Total Returns, by Quintile of Short-Term Past Returns, 1928-2025

However, the momentum portfolio doesn’t reliably regress to the mean: low returns over the past year and five years, in other words, don’t beget higher subsequent returns. Moreover, and unlike the Index, the momentum portfolio’s subsequent returns quickly erode: regardless of the quintile, comparing one column to the one on its right, returns in Table 5 erode (decrease) more markedly than in Table 4.

Most importantly, comparing each cell in Table 4 to the corresponding cell in Table 5, the former exceeds the latter. “Buy low” beats “buy high” because extreme returns at one point in time subsequently regress to their mean.

In short, momentum fails because it doesn’t merely ignore this fundamental reality: it vainly attempts to flout it.

My Results versus Others’

“For decades,” asserted The Australian Financial Review (9 January), “investors (sic) have generated solid returns from simply buying the previous year’s winners and selling the losers.”

That’s patently false. The so-called “trend” isn’t your friend: to the extent that it exists, it’s indistinguishable from random fluctuation.

Asness et al. contend that they “refute … myths (regarding momentum speculation).” They cite “academic papers (that have been presented and debated at top-level academic seminars and conferences, and have been published in peer-reviewed journals),” and analyse “the simplest data taken from Kenneth French’s publicly available website, a standard dataset used by both academics and practitioners.”

Asness et al. conclude that momentum speculation generates “an impressive long-term average return that survives all the attacks … hurled against it. Anyone repeating these myths, in any dimension, after reading (their article) is simply ignoring the facts.”

But what about the facts they ignore? The myths they perpetuate?

Among the facts they ignore: analysing the same momentum data which they utilised, it’s unarguable: a “buy and hold” investor’s portfolio, represented by the total, CPI-adjusted short- and medium-term returns of the S&P 500 Index, usually outperform– by a wide margin – the returns from momentum speculation. Cumulatively since 1927, the outperformance has been massive. In general, and ever more as the length of time increases, value emphatically trounces momentum.

Among the myths Asness et al. perpetuate: in their words, “the existence of momentum is a well-established empirical fact.” In truth, its existence in the U.S. over the past century is much more apparent than real; what they regard as “momentum” is, upon examination, indistinguishable from random fluctuation. I conclude that Asness et al. stand in the long queue of people who’ve been fooled by randomness.

They are “a little irked … by those who should know better but continue to repeat these myths, stretching the limits of credulity.” I’m amazed that they didn’t bother to test their assumptions’ veracity. I have – and my results render their and Faber’s conclusions untenable.

Implications

How Randomness Fools Momentum Speculators

Larry Swedroe (Rational Investing in Irrational Times: How to Avoid the Costly Mistakes Even Smart People Make Today, Talley, 2002) has illustrated how “runs” – and, by implication, “trends” and “momentum” – often reflect mere random fluctuation. Each year a statistics professor begins her class by asking all but one student to record the sequential outcomes (“H” for heads and “T” for tails) of 100 imaginary tosses of a coin. She also requests that one of the students toss a real coin and record the actual outcomes. The professor then leaves the room, lets the students conduct the tosses, imaginary and real, and returns later to assess the results.

Waiting on her lectern are 30 sheets of paper: each is numbered sequentially, from 1 to 30, and each contains a 100-character sequence of Hs and Ts. One records the actual tosses of a coin; 29 record students’ imagined tosses. The professor then announces that she’ll be able to distinguish the real sequence from the 29 imagined ones. After examining the 30 sequences, she amazes the class by choosing the correct one.

How did she do it? She knows that the sequence with the longest consecutive streak of Hs or Ts is the most likely to result from the toss of the real coin. She knows, in other words, that the sequence which contains the greatest “momentum” is actually likely to reflect random chance!

Given ten tosses of a fair coin, which sequence is more likely to occur: TTHHHHHHHH or HHTTHTHTTH? The answer, of course, is that each is equiprobable; yet psychologists have repeatedly found that virtually all people regard the second sequence as “more random” than the first. Hence none of the 29 students’ hypothetical sequences resembled the first sequence, and all of them resembled the second one.

Moreover, if H denotes “hit” and T denotes “miss,” and the sequences denote the success or failure of a shooter’s ten most recent shots, basketball players and fans will infer a “hot hand” from the first sequence. And if “T” denotes “negative return during the month” and H denotes “positive return during the month,” momentum speculators will infer that the first sequence reflects a stock which has developed “upward momentum.”

Note the similarity between the “hot hand” and the “momentum stock.” In each case the inference is highly questionable; players, fans and speculators have likely been fooled by randomness.

Forty years ago, Thomas Gilovich, Robert Vallone and Amos Tversky investigated the origin and validity of common beliefs regarding “streak shooting” (see “The hot hand in basketball: On the misperception of random sequences,” Cognitive Psychology, vol. 17, no. 3, July 1985). Most players and fans believe that the likelihood that a player’s next shot will be a “hit” is greater following a hit than a miss on the previous shot; they also believe that a player’s chance of scoring next time steadily rises as the number of consecutive previous “hits” increases.

However, Gilovich et al.’s “detailed analyses of … shooting records … provided no evidence of a positive correlation between the outcomes of successive shots.” The probability of a “hit” on the next shot is thus independent of the number of “hits” on recent shots. They attributed the “belief in the hot hand and the ‘detection’ of streaks in random sequences … to a general (misconception and underestimation) of chance …”

The New York Times devoted considerable attention to this research – which, Gary Belsky and Gilovich recounted in Why Smart People Make Big Money Mistakes: Lessons from the Life-Changing Science of Behavioral Economics (Simon & Schuster, 2009), “generated heated opposition … People were, and still are, unwilling to believe that the hot hand is a myth … The problems you might have in understanding the myth of the hot hand reflect the difficulty most people have with probability and statistics.”

Just as basketball players and fans reject the reality that the “hot hand” is a myth, momentum speculators refuse to accept that stocks’ and markets’ “momentum” is indistinguishable from random fluctuation.

In Malkiel’s words, “aside from the long-term positive direction of the stock market, streaks of excessively high returns don’t persist; indeed, lower returns typically follow them.” Overall markets’ short-term and medium-term returns tend to revert to their long-term means. The same is true, albeit less systematically, of the prices of individual companies’ stocks.

Not only did Icarus plummet to earth: Phoenix rose from the ashes. What soars usually crashes; yet what crashes often recovers. Hence “buy low and sell high” usually outperforms “buy high and sell even higher.”

Sometimes Crowds Are Wise; At Other Times They’re Crazy

Groups often make better decisions than individuals. If valid and reliable information is widely available and differing points of view are considered, discussion and debate among members of a group – which exposes and removes from consideration falsehoods, irrelevancies, etc. – help to produce better decisions. The system of free enterprise, widely dispersed ownership of private property, which produces market prices of assets, goods and services, is perhaps the best example of this “wisdom of crowds.”

If, however, information isn’t available, or isn’t regarded as valid and reliable, people are even more likely to heed alleged “experts” rather than think for themselves.

This paucity of information – particularly of historical “base rates” – occurs when people evaluate something new, like railways in the 19^th century, the Internet in the 1990s and AI today. Hence conformity often results from uncertainty: we follow others who assert that they know what they’re doing – even if we suspect that they don’t – because we don’t know that we don’t know. Moreover, the greater is the uncertainty and the higher are the stakes, the more vulnerable people become to the cues which foment herd behaviour.

Under these conditions, the wisdom of crowds wanes and the madness of crowds often reigns. At these junctures, people acting in markets and at the urgings of “experts” often make what in retrospect are clearly poor (at best) and disastrous (at worst) decisions.

When actors in markets defer to prominent, vocal and confident people, individuals’ biases and errors no longer cancel each other: instead, the biases and errors of alleged experts reinforce one another. “On Wall Street,” observe Belsky and Gilovich, “they call this investing with the herd, and the pervasiveness of this approach is expressed in the aphorism ‘the trend is your friend.’”

“Surely,” reckons Malkiel, “the wildly overoptimistic forecasts regarding the earnings potential of the Internet and the incorrect pricing of New Economy stocks … are examples of the pathology of herd behaviour.”

An economist, Sushil Bikhchandani, and two professors of finance, David Hirshleifer and Ivo Welch, coined the phrase “information cascade” to describe this phenomenon (see “A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades,” Journal of Political Economy, vol. 100, no. 5, October 1992). Essentially, they posit that trends (imagined or real) begin when individuals cease to think for themselves or rely upon their own information; instead, they focus upon the actions of others – even when they know that those actions contradict their own information and interests.

Powerful information cascades can lead prudent investors to become mindless speculators, that is, to buy simply because other people are buying, and sell simply because others are selling.

As Robert Shiller observed in Irrational Exuberance (Princeton University Press, 2015), information cascades and the madness of crowds feed upon themselves; specifically, they create “positive feedback loops.” An initial price rise encourages more people to buy, which in turn produces greater gains and induces even more people to buy. The phenomenon is an example of a Ponzi scheme and the “greater fool theory.”

“Eventually,” however, concludes Malkiel, “one runs out of greater fools.” “Any investment that has become a topic of widespread conversation,” he cautions, “is likely to be hazardous to your health … Inevitably, the hottest stocks or funds in one period are the worst performers in the next.”

It’s not just individual speculators who succumb to herding. Professionals, whose career risk trumps market risk and who mistakenly regard themselves as (or falsely claim they are) investors, stampede into the same stocks as amateurs. After all, pros are allegedly leaders: they must therefore follow the mob!

The “hottest” stocks, championed by the most exuberant “experts,” exhibit the greatest “momentum.” Momentum speculators, who chase these stocks, are thus members of the stampeding herd.

Herds stampede in both directions: just as they induce speculators to take – indeed, seek – excessive risk during periods of euphoria, they also prompt them to become overly risk-averse when the tide turns and pessimism prevails. Momentum speculators, like other members of the herd, steadily pour their money into “hot” stocks and funds when they’re rising and everyone is enthusiastic and confident – and hastily withdraw it when what was hot turns cold, prices fall and fear and pessimism reign.

Whether it’s on the upside or downside, Malkiel observes that “the media tend to encourage such self-destructive behaviour by hyping the severity of market (surges and plunges) and blowing events out of proportion to gain viewers and listeners.” Partly because of their poor timing, speculators typically earn a rate of return which is well below the one they’d earn by simply buying low and holding.

Never mind so-called “strong-form efficiency:” numerous studies have documented market participants’ tendency to over- and under-estimate the value of companies on a day-to-day, month-to-month and even year-to-year basis. Partly as a result of information cascades and herd behaviour – when they’re uncertain, people are prone to defer to the crowd – they often overreact to both good and bad news. As a result, the prices of popular companies rise too high and of unpopular ones fall too low.

Werner de Bondt and Richard Thaler’s landmark study – which provides a pillar of Leithner & Company’s operations – describes and explains this phenomenon (see “Does the Stock Market Overreact?” The Journal of Finance, vol. 40, no. 3, July 1985).

They analysed the performance of NYSE-listed stocks that had either risen or fallen more than the market average. They examined six-year and ten-year blocks of time, which they divided in half. Then, using returns from the first half of each period (which they dubbed the “formation period”) they constructed portfolios of “winners” (stocks whose gains were above-average) and losers (whose losses were above-average). Finally, they examined how the winners and losers performed over the second half of the study periods (“holding period”).

They found that “extreme returns of stocks (in the formation period) … were … subsequently followed (in the holding period) by significant price movement in the opposite direction. Using ten-year blocks of time, loser portfolios (generated) an average of 30% more than winner portfolios. Using the same procedure with six-year blocks of time, losers outperformed winners by almost 25% during the three-year holding periods.”

Notice the basic similarity of method and result between De Bondt and Thaler’s analysis and mine (Table 4 and Table 5).

They demonstrated what I’ve corroborated: when market participants react to extremes at one point in time, their reactions subsequently tend to reverse over time. De Bondt and Thaler established, in effect, that information cascades have two effects: they provide opportunities to generate losses as well as gains. When pessimism tamps a stock’s price, it eventually tends to recover; conversely, when exuberant buyers inflate a stock’s price, it later tends to recede (or collapse).

Hence speculators who buy high and expect to sell higher tend to underperform and even lose money; and investors who buy low and sell high outperform.

Both phenomena are examples of regression to the mean. They’re also a reminder that the crowd isn’t merely often wrong: it’s systematically wrong. Momentum speculation crashes into this fundamental reality; value investing successfully navigates it. Over short periods, buyers and sellers in financial markets can become deluded and misguided; but over the long run, value emerges – and value investors profit. That’s why Warren Buffett rejects momentum speculation. Quite the contrary: he famously urges that investors be “fearful when others are greedy and greedy when others are fearful.”

The best buying opportunities arise when the market participants are pessimistic and prices are plunging – not when they’re optimistic and prices are high and rising rapidly. For this reason, “buy low and sell high” usually outperforms “buy high and sell at even higher prices.”

Conclusion

In principle, momentum speculators strive to buy high and sell higher; in practice, they buy high and sell lower. They seek to exploit market inefficiencies and behavioral biases such as herd mentality, under- and over-reaction to news, and the fear of missing out (FOMO).

In fact, these phenomena defeat momentum speculators; they’ve succumbed to greed, followed the herd and been fooled by randomness.

Financial markets, Buffett observes, transfer wealth from the impatient (speculators) to the patient (investors). On 22 April 2025 he told CNBC: you “shouldn’t own stocks if you do dumb things.” By “dumb things” he meant buying a stock simply because its price has risen sharply, or selling it merely because its price has plunged. The implication (bearing in mind his view that “the dumbest reason in the world to buy a stock is because it’s going up”) is inescapable: by definition, momentum speculators do dumb things; therefore they shouldn’t own stocks!

Momentum speculation is a recipe for underperformance and loss. Value investing, in contrast, rests upon sound logical and empirical foundations. Ultimately, that’s why value investors typically trounce momentum speculators.

Click LIKE so that Livewire knows that you want more of this type of content. Click FOLLOW for notification when my next wire appears.