Momentum Investing Research: 30 Years of Academic Findings

Momentum is one of the most-studied phenomena in modern finance. Across three decades of peer-reviewed research, dozens of papers have tested, extended, challenged, and refined the original finding. What follows is a working reference: the papers that defined the field, what each one actually said, why it mattered, and where the field has gone since.

This is a long read by design. It’s meant to be cited, scanned, and returned to — the kind of resource you bookmark and come back to when you need the original source on a specific claim about momentum.

The short version. Momentum was formally documented in 1993, added to the standard four-factor model in 1997, conceded as “the premier anomaly” by the founders of efficient markets in 2008, and confirmed across nearly every asset class and centuries of price data in the years since. The literature has spent the last decade focused not on whether momentum works, but on how to manage its specific risks — particularly its tendency to crash during sharp market rebounds.

The foundational papers

Jegadeesh & Titman (1993) — “Returns to Buying Winners and Selling Losers”

The paper that started it all. Narasimhan Jegadeesh and Sheridan Titman, then at UCLA and the University of Maryland, examined U.S. stock returns from 1965 to 1989 and asked a simple question: does past performance predict future performance over an intermediate horizon?

The answer, in their data, was unambiguously yes. Stocks sorted into deciles by trailing 3-to-12-month returns continued to outperform or underperform over the subsequent 3 to 12 months. The top-decile (winners) minus bottom-decile (losers) strategy produced about 1% per month in excess return — an effect that was statistically robust and not explained by the standard risk models of the day.

This finding was a direct challenge to the efficient-markets hypothesis. If past returns predicted future returns, markets were not pricing information as efficiently as the theory required. The paper is now one of the most-cited works in financial economics, and the strategy it described — long winners, short or avoid losers, rebalance on a fixed schedule — is the canonical implementation of academic momentum.

Why it mattered: Formalized momentum as a measurable, statistically significant cross-sectional effect, and forced the academic mainstream to take it seriously as a real anomaly rather than data-mining noise.

Carhart (1997) — “On Persistence in Mutual Fund Performance”

Four years after Jegadeesh and Titman, Mark Carhart asked a related question: how much of the apparent skill in mutual fund returns is actually just momentum exposure?

His answer reshaped how the field evaluates investment performance. Carhart added a momentum factor to the existing Fama-French three-factor model (market, size, value), creating the four-factor Carhart model. When applied to the universe of mutual funds, the new model showed that much of what looked like manager skill was, in fact, persistent exposure to momentum. Funds that appeared to outperform were largely riding the momentum factor; very few showed evidence of true alpha after controlling for it.

The model itself became the workhorse for fund performance attribution. If you’ve ever read a paper that controls for “the four factors,” it’s Carhart’s model.

Why it mattered: Made momentum a standard part of the asset-pricing toolkit. It is no longer possible to evaluate a fund’s performance seriously without controlling for momentum exposure.

Fama & French (2008) — “Dissecting Anomalies”

Eugene Fama and Kenneth French are the architects of the efficient-markets framework that momentum awkwardly contradicts. So when they themselves examined the universe of documented anomalies and called momentum “the premier anomaly,” it carried unusual weight.

Their paper systematically tested a wide range of candidate anomalies against rigorous statistical and economic significance criteria. Most failed to hold up — they were either too weak, too small, too unstable, or too easily explained by known risk factors. Momentum was the standout exception. It survived every test, across every market segment they examined.

When the people whose theory an effect contradicts call it the strongest documented anomaly in finance, that’s about as strong an endorsement as the academic literature ever produces.

Why it mattered: Removed any remaining doubt — among the people best positioned to be skeptical — that momentum was a real, persistent, and important market phenomenon.

The expansion: momentum everywhere

Asness, Moskowitz & Pedersen (2013) — “Value and Momentum Everywhere”

For the first 20 years after Jegadeesh and Titman, most momentum research focused on U.S. stocks. Cliff Asness, Tobias Moskowitz, and Lasse Pedersen took the question much wider: does momentum exist outside of U.S. equities?

Their answer: yes, and with remarkable consistency. They documented momentum in U.S. equities, international equities, country index returns, government bonds, currencies, and commodities. Across six asset classes spanning more than 40 years of data, momentum produced statistically and economically significant returns in nearly every market they tested.

The paper also produced one of the most-cited findings in factor investing: value and momentum are negatively correlated across asset classes. A portfolio combining both factors produced a higher Sharpe ratio than either alone — a finding now treated as foundational in multi-factor portfolio construction. (See our companion piece: Value and Momentum: Why the Combination Beats Either Alone.)

Why it mattered: Generalized momentum from a stock-market anomaly to a near-universal feature of asset prices, and established the diversification case for combining it with value.

Moskowitz, Ooi & Pedersen (2012) — “Time Series Momentum”

A complementary line of research from the same year asked a slightly different question. The original momentum studies looked at cross-sectional momentum: ranking assets against each other within a universe. Moskowitz, Ooi, and Pedersen examined time-series momentum: whether an asset’s own past return predicts its own future return.

They found a strong, persistent effect across 58 liquid futures markets over four decades. An asset that had been rising over the past 12 months tended to continue rising; one that had been falling tended to continue falling. The effect was significant in every asset class and on every horizon they examined.

This finding has practical importance because time-series momentum is the foundation of the trend-following strategies used by many systematic managers. It connects the cross-sectional momentum of the equity literature to the broader systematic-trading universe.

Why it mattered: Established that momentum is not just a relative-strength phenomenon — it shows up in absolute terms as well, and applies to single-asset trend-following as much as to cross-sectional stock selection.

Geczy & Samonov — Two-Century Momentum Studies

Christopher Geczy and Mikhail Samonov extended the momentum record about as far back as financial data exists. Across a series of papers, they tested momentum strategies on U.S. stocks back to 1801, on global equities back to 1800, and on a broader universe spanning more than two centuries.

The finding: momentum has been a persistent feature of markets for as long as we have data to test it on. Even out-of-sample relative to all of the post-1993 research that might have priced it away, the effect persisted across the deep historical record.

Why it mattered: Foreclosed the criticism that momentum might be a recent artifact of modern market structure. If the effect was present in 19th-century markets, it isn’t a product of high-frequency trading, the rise of ETFs, or any modern institutional feature.

The risk-management literature

The decade since 2013 has shifted the research focus from whether momentum works to how to manage its specific risks. Two papers stand out.

Daniel & Moskowitz (2016) — “Momentum Crashes”

Kent Daniel and Tobias Moskowitz documented momentum’s worst-case behavior in detail. Examining U.S. equity momentum across nearly 90 years of data, they showed that the strategy’s worst losses cluster in a specific regime: sharp market rebounds following extended drawdowns.

When the market plunges and then snaps back, the most beaten-down losers tend to rocket higher fastest. A momentum strategy, positioned in recent winners and against those losers, gets caught on the wrong side. The 1932 and 2009 rebounds each produced momentum drawdowns of more than 70% in a matter of months.

The paper’s most important contribution wasn’t documenting the crashes — it was showing they were partially forecastable. The strategy’s own recent volatility and the market’s drawdown state contained information about the likelihood of a crash forming. That meant the risk could, at least in principle, be managed rather than simply borne.

For a deeper treatment of this failure mode and how to manage it, see Momentum Crashes: The Hidden Risk in Momentum Investing.

Why it mattered: Named, documented, and forecast a previously under-appreciated failure mode of momentum — turning crash risk from a buried tail into a manageable engineering problem.

Barroso & Santa-Clara (2015) — “Momentum Has Its Moments”

Pedro Barroso and Pedro Santa-Clara published the most influential answer to the question Daniel and Moskowitz had raised: if crashes are partially forecastable, what’s the best way to manage them?

Their proposal was elegant in its simplicity: scale the entire momentum portfolio’s exposure up or down based on its own recent realized volatility. When the strategy is calm, run more of it. When it’s turbulent, run less. The paper showed that this time-series volatility scaling materially improved the risk-adjusted return of momentum across a long historical sample — not by eliminating crashes entirely, but by being smaller when they hit.

The Barroso-Santa-Clara line of work has been replicated, extended, and applied across asset classes. Volatility-managed momentum is now a standard implementation choice in the academic and applied literatures.

Why it mattered: Provided the canonical mitigation for momentum crash risk — and a template for how systematic strategies more generally should think about scaling exposure based on observable risk signals.

The residual momentum line

A parallel strand of research has focused on how to construct a cleaner momentum signal.

Blitz, Huij & Martens (2011) — “Residual Momentum”

David Blitz, Joop Huij, and Martin Martens proposed an alternative: instead of ranking stocks by raw past return, rank them by the residual return — the part of each stock’s past performance not explained by the standard factor exposures (market, size, value).

The intuition: raw momentum implicitly bundles together a stock’s actual idiosyncratic trend and its drift caused by exposure to other factors. Residual momentum isolates the idiosyncratic piece. The empirical result was a momentum signal with comparable returns to standard momentum but a substantially lower drawdown profile, particularly during crash periods.

Residual momentum has become a common choice in academic and quantitative implementations seeking to harvest the momentum premium with less of the crash risk. It is conceptually adjacent to — but distinct from — the volatility-scaling approach.

Why it mattered: Showed that the momentum signal itself can be refined, not just the exposure to it. Where Barroso–Santa-Clara manages how much momentum to hold, Blitz–Huij–Martens manages what kind of momentum to hold.

The behavioral and economic explanations

The empirical evidence for momentum is overwhelming. The harder question — and the one the literature continues to debate — is why. Several explanations have substantial supporting research:

• Behavioral underreaction. Investors anchor to old prices and revise their views too slowly when new information arrives. Good news propagates over weeks or months rather than instantly, and momentum harvests the resulting drift. Daniel, Hirshleifer & Subrahmanyam (1998) and Hong & Stein (1999) developed influential models along these lines.
• Behavioral overreaction. Late in a trend, investors herd into recent winners and push prices beyond fundamentals. This explains the eventual reversal that follows momentum at very long horizons, and is part of the same behavioral picture.
• Risk compensation. Part of momentum’s return may be straightforward compensation for bearing crash risk. The strategy works most of the time but periodically suffers severe losses; rational investors would demand a premium for taking that profile on. This view is most consistent with the Daniel-Moskowitz crash findings.
• Institutional and capacity constraints. Some research argues momentum persists because the institutions best positioned to arbitrage it away face structural constraints — tracking-error budgets, position limits, drawdown sensitivity — that prevent them from taking it on at the scale needed to eliminate the effect.

These explanations are not mutually exclusive. The likely truth is that momentum reflects a combination of behavioral patterns, risk compensation, and institutional limits — which is part of why it has been so resistant to being arbitraged away.

Synthesis: what 30 years of research actually shows

The research record, taken as a whole, supports a small number of robust statements:

1. Momentum is real. The cross-sectional return effect documented by Jegadeesh and Titman has been replicated thousands of times, across markets, asset classes, and centuries.
2. It is the most-documented anomaly in finance — by the explicit judgment of its most credible skeptics.
3. It survived its own publication. Out-of-sample work since 1993 has confirmed the effect persists; if anything, the literature is more confident in it now than at the time of its discovery.
4. It is not free of risk. The strategy is vulnerable to specific, well-characterized crash periods, and naive implementations bear that risk in full.
5. Crash risk is partially manageable. Risk-aware ranking, volatility-based exposure scaling, and regime-sensitive allocation each reduce the worst drawdowns without giving up most of the long-run edge.
6. Combining momentum with negatively-correlated factors (notably value) produces a higher Sharpe ratio than either alone. This is one of the most robust findings in factor investing.

These six statements are the durable core of the literature. Almost everything else in the momentum research record is detail, extension, or refinement around them.

How SignalStrike treats this research

SignalStrike was built by people who have read this literature carefully and have deployed their own capital based on it. The platform reflects a few specific takeaways:

• The mechanics are configurable, not hidden. Universe, lookback, ranking method, weighting, rebalance schedule — all the parameters the research identifies as load-bearing are visible and adjustable. This is research-grade transparency, not a black-box “trust us” product.
• Risk management is a first-order feature, not an afterthought. Volatility-adjusted ranking and regime-responsive tools — including options to rotate toward bonds or cash — exist specifically because the crash-risk literature exists.
• You can test before you commit. Every configuration is backtestable against historical data, and the same engine runs backtests and the live screen. That means strategy choices can be evaluated on their own historical record before they’re trusted with real capital.

For RIAs and serious investors who want the full methodology walked through in detail, it is available. The founders’ public portfolio tracker shows how their own implementation has performed in live markets.

Frequently Asked Questions

What is the most important academic paper on momentum investing?

Jegadeesh and Titman’s 1993 paper “Returns to Buying Winners and Selling Losers” is the foundational work — it formalized momentum as a measurable cross-sectional effect and is among the most-cited papers in financial economics. For the modern toolkit, Carhart’s 1997 paper added momentum to the standard four-factor model, and Asness, Moskowitz, and Pedersen’s 2013 paper “Value and Momentum Everywhere” generalized the finding across asset classes.

Does momentum investing work according to academic research?

The academic record is unusually one-sided on this question. The effect has been documented in original samples, confirmed out-of-sample after publication, found across multiple asset classes, and traced back through more than two centuries of price data. Even Eugene Fama and Kenneth French — whose efficient-markets framework momentum contradicts — called it “the premier anomaly” in 2008.

What is the Carhart four-factor model?

The Carhart four-factor model adds a momentum factor to the standard Fama-French three-factor model (market, size, value), creating a four-factor framework for explaining stock returns. It was introduced by Mark Carhart in 1997 and has become the standard tool for evaluating fund performance — including in determining whether a fund’s apparent skill is actually just persistent momentum exposure.

What does the research say about momentum crashes?

Daniel and Moskowitz’s 2016 paper “Momentum Crashes” showed that momentum’s worst drawdowns tend to cluster in a specific regime: sharp market rebounds following extended bear markets. Crucially, the paper also showed these crash conditions are partially forecastable from observable inputs — most importantly the strategy’s own recent volatility — which is why active risk management is the standard response in the modern literature.

How long has momentum been documented in markets?

Geczy and Samonov tested momentum strategies on U.S. stocks back to 1801 and on global markets across more than two centuries of data. The effect was present in essentially every period and market they examined. This is a key piece of evidence that momentum is not a recent artifact of electronic trading or modern institutional structure — it appears to be a structural feature of how prices respond to information.

Further Reading

• Jegadeesh, N., & Titman, S. (1993). “Returns to Buying Winners and Selling Losers.” Journal of Finance.
• Jegadeesh, N., & Titman, S. (2001). “Profitability of Momentum Strategies.” Journal of Finance.
• Carhart, M. M. (1997). “On Persistence in Mutual Fund Performance.” Journal of Finance.
• Daniel, K., Hirshleifer, D., & Subrahmanyam, A. (1998). “Investor Psychology and Security Market Under- and Overreactions.” Journal of Finance.
• Hong, H., & Stein, J. C. (1999). “A Unified Theory of Underreaction, Momentum Trading, and Overreaction.” Journal of Finance.
• Fama, E. F., & French, K. R. (2008). “Dissecting Anomalies.” Journal of Finance.
• Blitz, D., Huij, J., & Martens, M. (2011). “Residual Momentum.” Journal of Empirical Finance.
• Moskowitz, T. J., Ooi, Y. H., & Pedersen, L. H. (2012). “Time Series Momentum.” Journal of Financial Economics.
• Asness, C. S., Moskowitz, T. J., & Pedersen, L. H. (2013). “Value and Momentum Everywhere.” Journal of Finance.
• Barroso, P., & Santa-Clara, P. (2015). “Momentum Has Its Moments.” Journal of Financial Economics.
• Daniel, K., & Moskowitz, T. J. (2016). “Momentum Crashes.” Journal of Financial Economics.
• For an accessible introduction to the topic, see our pillar overview: Momentum Investing: How It Works and Why It Persists.

Related Research

• What Is Momentum Investing? A Complete Guide — Learn how momentum investing works - the academic research, why baskets beat individual picks, and why monthly rebalancing outperforms trailing stops.
• Value and Momentum: Why the Combination Beats Either Alone — Value and momentum are negatively correlated — they shine and struggle at different times. Why combining them has produced steadier results than either alone.
• Momentum Crashes: The Hidden Risk in Momentum Investing — Momentum crashes are sudden, severe losses in momentum strategies — usually right after a panic. Here is when they occur, why, and how to manage them.

Disclosures