Momentum Investing: How Past Winners Keep Winning

Momentum investing challenges one of the most intuitive assumptions in markets: that what goes up must come down. In reality, decades of academic research — beginning with Jegadeesh and Titman’s landmark 1993 study — show that stocks with strong recent performance tend to keep outperforming over the following 3 to 12 months, while recent losers tend to keep underperforming.

This momentum effect is one of the most robust anomalies in financial economics. It has been documented across U.S. equities, international markets (Europe, Asia, emerging economies), and even across asset classes including bonds, commodities, and currencies — with evidence spanning more than a century. This guide covers what momentum investing is, how it works, why it persists, and the significant risks that come with it.

What is Momentum Investing?

Momentum investing is a systematic strategy that buys securities with strong recent performance and sells (or avoids) securities with weak recent performance. Unlike speculative trend-chasing, momentum investing follows a disciplined, rules-based framework grounded in rigorous academic evidence.

The academic foundation rests on two landmark papers. Jegadeesh and Titman (1993) first documented that buying past winners and selling past losers within the U.S. stock market produced statistically significant profits over 3- to 12-month horizons. Mark Carhart (1997) then formalized momentum as the fourth factor in an extension of the Fama-French Three-Factor Model, demonstrating that momentum explained a significant portion of mutual fund performance persistence.

An important nuance: momentum operates at an intermediate horizon. At very short horizons (days to weeks), stock returns tend to exhibit reversal — recent winners pull back slightly. At very long horizons (3 to 5 years), returns also tend to mean-revert, as overvalued winners eventually correct. Momentum captures the sweet spot in between, where trends persist before eventually fading.

Types of Momentum

Researchers and practitioners distinguish three main types of momentum, each with different construction methods and applications:

1. Cross-Sectional Momentum (Relative Strength) — The most studied form. Stocks are ranked against each other within a universe (e.g., the S&P 500), and the strategy goes long the top performers and short the bottom performers. The key insight is relative performance: which stocks are winning compared to their peers, regardless of whether the overall market is up or down.

2. Time-Series Momentum (Trend-Following) — Each asset is evaluated against its own history, not relative to peers. If a stock’s recent return is positive, go long; if negative, go short or stay flat. This approach is the foundation of commodity trading advisor (CTA) and managed futures strategies.

3. Industry Momentum — Moskowitz and Grinblatt (1999) showed that rotating into outperforming industries and out of underperforming ones captures a significant portion of individual stock momentum. Sector-level momentum tends to be less noisy than individual stock signals.

Most academic research and factor-based ETFs use cross-sectional momentum with the 12-1 signal as the standard implementation.

The Momentum Factor

In factor models, the momentum premium is captured by a long-short portfolio that isolates the return difference between winners and losers:

The factor goes by two common names: UMD (Up Minus Down) and WML (Winners Minus Losers). Both refer to the same construction — the return spread between the best- and worst-performing stocks ranked by past returns.

Carhart (1997) added UMD as the fourth factor to the Fama-French three-factor model, creating the four-factor model widely used in performance attribution: market risk premium, SMB (size), HML (value), and UMD (momentum). For the full three-factor framework and how momentum extends it, see our guide to the Fama-French Three-Factor Model.

The long-run momentum premium in U.S. equities has averaged approximately 6-8% annually on a gross long-short basis (before transaction costs and implementation frictions), though this figure varies significantly by time period, universe, and construction methodology. Critically, the premium has been confirmed across international equity markets, government bonds, corporate bonds, commodities, and currencies — making momentum one of the most pervasive factors in financial economics.

Momentum Investing Example

The example highlights two realities of momentum investing: the gross premium can be substantial, but transaction costs meaningfully erode returns because the strategy requires frequent portfolio turnover as momentum signals shift.

Behavioral Explanations for Momentum

Why does momentum persist despite being widely documented for decades? The leading explanations come from behavioral finance, which identifies systematic cognitive biases that cause investors to react slowly to new information:

1. Underreaction (Anchoring and Conservatism) — When a company reports strong earnings or announces positive news, investors anchored to prior expectations only partially adjust their valuations. This slow belief updating creates a gradual price drift in the direction of the news — the momentum effect.

2. Herding and Trend Reinforcement — Once a price trend becomes visible, more investors pile in, creating positive feedback loops. Rising prices attract attention, media coverage, and additional capital, pushing prices further in the same direction before fundamentals catch up.

3. Overconfidence and Delayed Diffusion — Investors in winning positions tend to become overconfident and increase exposure, while bad news about losing stocks diffuses slowly — particularly for smaller, less-followed companies where fewer analysts provide coverage.

Momentum vs Value Investing

Momentum and value are two of the most studied factor strategies in finance, and they operate on fundamentally different principles:

One of the most useful properties for portfolio construction is that momentum and value have exhibited historically low or negative correlation on average, though this relationship is time-varying. When momentum outperforms, value tends to underperform — and vice versa. This makes combining both factors an effective diversification strategy. For a comprehensive comparison of investment styles, see our guide to growth vs value investing.

Momentum Crashes

The most severe momentum crash in modern history occurred in 2009. After the 2008 financial crisis, momentum portfolios were positioned long in defensive stocks (which had outperformed during the crash) and short in beaten-down financials and cyclicals. When the market recovery began in March 2009, the losers rebounded sharply — some of the most battered stocks surged dramatically — while the defensive winners stagnated. The long-short momentum portfolio suffered its worst drawdown on record.

Why crashes happen: Momentum crashes occur during convex market rebounds — when the market quickly reverses direction after a sharp decline. The strategy is effectively short the most beaten-down stocks, which are precisely the ones that rebound most aggressively when sentiment shifts. The short leg of the portfolio amplifies losses during these reversals; long-only momentum implementations suffer less but are still affected. For more on the mechanics and risks of the short side, see our guide to short selling.

Statistically, momentum returns exhibit negative skewness (occasional extreme left-tail events) and positive kurtosis (fat tails). This tail risk profile makes momentum fundamentally unsuitable as a standalone strategy and underscores the importance of combining it with other factors.

How to Analyze Momentum

Implementing momentum in practice requires careful attention to construction details and costs:

1. Choose long-only or long-short: Most individual investors use long-only momentum — buying the top-performing stocks while avoiding or underweighting the losers. The full long-short strategy adds the short leg (selling losers), which amplifies returns but introduces borrowing costs, recall risk, and significantly higher crash exposure.
2. Define the signal: The standard 12-1 signal (past 12-month return excluding the most recent month) is the most common. Some implementations use shorter formation periods (6-1) or combine multiple lookback windows.
3. Set a rebalance cadence: Monthly or quarterly rebalancing is typical. Momentum signals decay over time, so portfolios must be refreshed regularly — but more frequent rebalancing increases transaction costs.
4. Monitor transaction costs and tax drag: Momentum has the highest turnover of any standard factor strategy. Track estimated trading costs (bid-ask spreads, market impact) and consider tax-loss harvesting opportunities on the losing positions.
5. Combine with other factors: Pair momentum with value, quality, and low-volatility exposures. These factors have historically low correlation with momentum and help offset its crash risk.

Common Mistakes

1. Confusing momentum with speculation. Momentum investing is a systematic, rules-based strategy grounded in decades of peer-reviewed academic research. It is not “buying high and hoping it goes higher.” The distinction lies in defined formation periods, portfolio construction rules, diversification across dozens or hundreds of stocks, and disciplined rebalancing.

2. Ignoring transaction costs. Momentum has the highest turnover of any major factor strategy. The theoretical premium of 6-8% annually can be significantly eroded — sometimes by half or more — by real-world trading costs, bid-ask spreads, market impact, and tax drag from short-term capital gains.

3. Not accounting for crash risk. Treating momentum as a “free lunch” ignores the strategy’s severe tail risk. Momentum crashes — like the 2009 episode — can wipe out years of accumulated gains in weeks. Always size momentum as one component of a diversified multi-factor portfolio, not as a standalone allocation.

4. Applying momentum to individual stocks without diversification. Momentum signals are noisy for any single stock. The strategy works because the signal is consistent in aggregate across hundreds of stocks. Running a concentrated 5- or 10-stock “momentum portfolio” is not momentum investing — it is concentrated speculation with momentum characteristics.

5. Backtest errors. Momentum strategies are particularly susceptible to look-ahead bias (using information not available at the time of the trade), survivorship bias (excluding delisted stocks from the backtest universe), and data-snooping (overfitting to historical patterns). Any momentum backtest should account for realistic transaction costs, use point-in-time data, and include delisted securities.

Limitations of Momentum Investing

1. High turnover and transaction costs. Monthly or quarterly rebalancing generates substantial trading frictions. The theoretical momentum premium is meaningfully eroded by bid-ask spreads, market impact costs, and — for the short leg — stock borrowing fees.

2. Tax inefficiency. Frequent trading creates short-term capital gains, which are taxed at ordinary income rates rather than the lower long-term capital gains rate. Momentum is one of the least tax-efficient factor strategies.

3. Crash risk. As discussed above, momentum returns exhibit negative skewness and excess kurtosis. The strategy’s tail risk is fundamentally different from — and more severe than — the gradual underperformance that value investors occasionally experience.

4. Capacity constraints. Momentum strategies perform worse at scale. As more capital chases the same momentum signals, the price impact of trading erodes the premium. This is a significant concern for large institutional investors.

5. Short leg impracticality. The full academic momentum strategy requires short-selling the losers portfolio, which involves borrowing costs, recall risk, and regulatory constraints. Most retail investors and many institutional investors implement long-only variants, which capture only a portion of the theoretical premium.