Cross-sectional momentum
MomentumIntermediate
Cross-sectional momentum ranks a universe of stocks by their recent returns and owns the winners relative to the losers. The classic research result is that stocks which outperformed over the past 3 to 12 months tend to keep outperforming over a similar horizon. A common implementation skips the most recent month, because that month often reverses.
Not investment advice. Nothing here is a recommendation to buy or sell anything. Backtests and historical studies are not guarantees, and a rule that worked in one sample can fail in the next. Trading can lose money. Links are for education, not endorsements. Do your own research.
- Rules
- Each month, rank stocks by return from 12 months ago to 1 month ago. Buy the top decile or quintile and, in the long-short version, short the bottom. Hold for a month to a few months, and rebalance. Long-only investors can simply overweight the winners inside an equity book. Exclude illiquid names and very low prices, or the result is a small-cap trading strategy in disguise.
- Where AI fits
- AI is useful for the universe screen, the skip-month choice, and a risk model that stops the portfolio from being a hidden bet on one sector. Machine-learning papers often find momentum-like features among the signals that survive, which is a reason to treat this as an input to a model rather than something a model should replace blindly. Rebalancing and borrow costs have to be in the test.
- Markets
- Liquid stocks in the US and other developed markets. The same ranking idea is applied to sectors and asset classes.
- Timeframes
- Monthly rebalance using daily or monthly returns.
- Risks
- Momentum has sharp crashes when last year's losers rebound together. The short side is expensive and sometimes unavailable. Capacity is limited in small caps, and a backtest without borrow fees and spreads overstates the short book.