Cross-sectional and time-series momentum
Two distinct momentum strategies exist. Cross-sectional momentum is relative: rank all stocks by past returns, go long winners and short losers. Returns come from relative outperformance, not market direction — it is market-neutral in design. Time-series (or trend-following) momentum is absolute: go long an asset if it has been rising (positive past return relative to its own history), short if falling. Time-series momentum underpins managed futures / CTA strategies and has shown strong performance in crisis periods — when it tends to be long short positions and short equities during crashes, providing crisis alpha that complements equity portfolios.
Why momentum exists: competing explanations
The persistence of momentum has generated extensive theoretical debate. Behavioural explanations are the most compelling: under-reaction to news (investors are slow to update their views on a company’s improving fundamentals, allowing price to drift upward over months), investor herding, and the disposition effect (investors hold losers too long and sell winners too early, creating continuation patterns). Risk-based explanations are less convincing — momentum strategies crash severely (2009 momentum crash: −73%), suggesting the risk characteristics are unattractive rather than compensated. Daniel and Moskowitz (2016) showed that momentum has "crash risk" — it does poorly precisely when the market rebounds sharply from a crisis bottom.
The momentum crash problem
Momentum strategies are vulnerable to sharp reversals — "momentum crashes" — that tend to occur at the beginning of sharp market recoveries. During 2008, momentum was long quality/defensive stocks and short financials (which had fallen most). When markets rebounded in March 2009, the short portfolio (beaten-down stocks) recovered most violently, causing momentum losses of 40–70% in a few weeks. This crash risk means momentum cannot simply be mixed into a portfolio without explicit crash protection: dynamic position sizing that scales down momentum exposure in periods of high market volatility (the Daniel-Moskowitz signal) substantially improves its risk-adjusted performance.
“Momentum is the embarrassment of efficient market theory. It exists, it has persisted, it is robust across markets and time — and we still cannot fully agree why.” — Eugene Fama
What this means for you
Momentum investing is available to retail investors through dedicated momentum ETFs (such as iShares Edge MSCI USA Momentum Factor ETF) and through systematic trend-following funds. The practical challenge is that raw momentum strategies have significant transaction costs (high turnover) and crash risk, making professional implementation preferable to naive replication. From a portfolio construction perspective, momentum combined with value provides a compelling combination — they are negatively correlated factors, meaning momentum tends to outperform when value underperforms and vice versa, reducing portfolio drawdowns when blended. AQR’s evidence on value + momentum blends remains the most compelling practical case for combining these factors systematically.