What Is Momentum Investing? The Academic Research Behind the Strategy

Most serious investing strategies can be traced back to a simple, defensible idea. Value investing asks: am I paying less than something is worth? Quality investing asks: am I buying a business with durable competitive advantages? Momentum investing asks a different question entirely: is this asset moving in a direction that is likely to continue?

That question sounds like market timing or trend-chasing — the kind of thing academic finance has traditionally dismissed. But the evidence tells a different story. Momentum has been studied, replicated, and stress-tested across markets, time periods, and asset classes for over 30 years. It is not a fluke.

The Core Idea

Momentum investing rests on a straightforward empirical observation: assets that have performed well over the recent past tend to continue performing well over the near future, and assets that have performed poorly tend to continue underperforming. You buy the recent winners and avoid — or short — the recent losers.

The "recent past" that matters is typically the prior 3 to 12 months of returns. Crucially, practitioners usually skip the most recent month when forming the signal, because very short-term returns (one month and under) exhibit a reversal effect rather than continuation — a separate phenomenon driven by microstructure and liquidity effects.

The holding period that captures momentum most cleanly is somewhere between one and twelve months. Shorter than that, you're picking up noise and reversal. Longer than that, you start picking up value effects running in the opposite direction.

This is not a vague qualitative observation. It has been quantified, replicated, and subjected to out-of-sample tests across half a century of data in dozens of markets around the world.

The Landmark Paper: Jegadeesh & Titman (1993)

The study that put momentum on the map for academic finance was published in the Journal of Finance in 1993 by Narasimhan Jegadeesh and Sheridan Titman. Working with U.S. equities from 1965 to 1989, they sorted stocks into deciles based on their past 3-to-12-month returns and examined what happened to those groups over the following 3 to 12 months.

The finding was striking: portfolios of recent winners consistently outperformed portfolios of recent losers by meaningful margins. The spread between the top and bottom deciles — a long-short portfolio that buys the best performers and sells the worst — generated substantial risk-adjusted returns that could not be explained by standard measures of market risk.

What made the paper especially influential was its rigor. The authors controlled for size, bid-ask spreads, and short-term reversal. The momentum effect was not an artifact of data mining or microstructure noise. It was real, large, and persistent throughout the sample. Subsequent research has confirmed the same pattern in U.S. equities out-of-sample through the 1990s, 2000s, 2010s, and beyond.

Why Does It Persist? Behavioral vs. Risk-Based Explanations

If momentum is real and well-documented, a natural question follows: why hasn't it been arbitraged away? Two broad families of explanations compete for the answer.

Behavioral explanations

The more widely accepted view in academia holds that momentum persists because of predictable investor psychology. Three mechanisms stand out:

• Underreaction to news. When a company reports genuinely good earnings or a currency is supported by improving fundamentals, investors update their views slowly. Prices drift upward for months rather than instantly reflecting all available information. Early movers profit; late movers keep pushing prices in the same direction.
• Herding. Institutional investors and retail traders alike are influenced by what others are doing. A rising price attracts attention, media coverage, and new capital — creating a self-reinforcing cycle that can outlast the original fundamental catalyst.
• The disposition effect. Investors tend to sell winners too quickly (locking in gains) and hold losers too long (avoiding the pain of realizing a loss). This behavior suppresses upward price movement in winners and slows downward price movement in losers, creating a persistent lag that momentum strategies can exploit.

Risk-based explanations

A minority view holds that momentum returns compensate for some form of risk that is not fully captured by standard models. Perhaps momentum portfolios are exposed to crash risk (they perform very badly in sudden market reversals), or they co-vary with business cycle risk in ways that make their returns genuinely hard to earn. This view has some empirical support — momentum does experience severe drawdowns in specific environments — but most researchers consider it insufficient to explain the full magnitude of the momentum premium.

The honest answer is that both explanations probably contribute. Behavioral biases create the initial inefficiency; risk and transaction costs prevent it from being fully arbitraged out.

Momentum Across Asset Classes

A finding that only works in U.S. equities could plausibly be a statistical artifact. A finding that works in stocks, bonds, currencies, and commodities simultaneously — across 40+ countries and several decades — is much harder to dismiss.

That is exactly what Asness, Moskowitz, and Pedersen demonstrated in their 2013 Journal of Finance paper "Value and Momentum Everywhere." They examined momentum (and value) signals across eight distinct asset classes: U.S. equities, U.K. equities, European equities, Japanese equities, equity index futures, government bond futures, currency forwards, and commodity futures.

The cross-asset finding matters for several reasons. First, it strongly argues against data mining — the same pattern does not accidentally appear across all these markets by chance. Second, it shows that momentum is not a story about any particular market's microstructure or investor base. Third, it opens the door to building genuinely diversified momentum portfolios that span multiple asset classes simultaneously.

How R2S Implements Momentum

R2S runs two distinct momentum strategies, each capturing a different dimension of the effect.

TSMOM — Time-Series Momentum

Time-series momentum (TSMOM) asks a single question about each instrument in isolation: is this asset trending up or down relative to its own history? The signal is computed by comparing an instrument's current price to its price 12 months ago, adjusted for recent volatility to normalize position sizing across assets with very different return distributions.

A TSMOM signal is long when the instrument has delivered positive returns over the lookback window and flat (or short, where applicable) when it has not. It does not require comparison with other assets — an instrument can be a TSMOM buy even if every other instrument in its category is falling, as long as it is up on an absolute basis.

XS_MOM — Cross-Sectional Momentum

Cross-sectional momentum (XS_MOM) ranks instruments against each other. Within a given universe — equities, commodities, currencies — assets are sorted by their recent returns, and the strategy takes long positions in the relative winners and avoids or shorts the relative losers. An instrument's absolute performance does not matter; what matters is whether it is beating its peers.

The two approaches are complementary. TSMOM captures absolute trend; XS_MOM captures relative strength. They tend to diverge during broad market rallies or crashes, where assets move together, and converge when dispersion across instruments is high. Running both provides more stable signal coverage across different market regimes.

Practical Considerations: Crashes, Costs, and Portfolio Context

No honest account of momentum investing is complete without addressing its failure modes. Momentum is not a free lunch.

Momentum crashes. The most well-documented risk is the momentum crash — a sudden, severe reversal that occurs when a prolonged trend abruptly ends, usually during or immediately after a market crisis. In these environments, the recent winners that a momentum portfolio is long often turn out to be the most overextended assets, and the losers being avoided or shorted experience sharp rebounds. The drawdowns can be large and fast. March and April 2009 are the canonical example: as markets bounced off the financial crisis lows, prior losers (beaten-down financials, cyclicals) surged while prior winners fell sharply.

Transaction costs. Momentum strategies turn over their portfolios more frequently than value or quality strategies, which means transaction costs matter. Monthly rebalancing at the instrument level is a deliberate design choice at R2S to balance signal freshness against turnover. This is a meaningful advantage relative to daily or weekly rebalancing approaches that can significantly erode gross returns after costs.

Diversification benefits. Despite its crash risk, momentum has a low or negative correlation with value strategies, which tend to perform best precisely when momentum struggles — in sharp reversals. This makes momentum a strong complement to other factor exposures rather than a standalone strategy. A portfolio combining value, momentum, and quality factors achieves better risk-adjusted returns than any single factor alone, a result confirmed across multiple decades and geographies.

The diversification benefit of combining momentum with other factors is not a coincidence. Value and momentum tend to be negatively correlated because they respond to the same environment — trend continuation vs. mean reversion — in opposite ways.
— Concept from Asness, Moskowitz & Pedersen (2013)

This is why R2S presents momentum signals alongside other strategy signals on the dashboard. The goal is not to run momentum in isolation but to help subscribers understand how momentum exposure fits within a broader, diversified factor framework.

The Bottom Line

Momentum is one of the rare investment strategies that has survived every empirical test thrown at it: out-of-sample validation, international replication, multiple asset classes, and decades of post-publication live data. It is not perfect — it has real crash risk and requires disciplined portfolio construction to harvest effectively. But the underlying phenomenon is as well-documented as anything in empirical finance.

Understanding the academic foundation behind a strategy matters. When momentum underperforms for a year or two — and it will — investors who understand why momentum works, and why it sometimes does not, are far more likely to stay disciplined than those who adopted it purely on the basis of backtested results.