Strategy #0000000013 · Cross-sectional momentum

Price momentum
Rank assets by past returns, then buy winners and sell losers.

At each formation date, rank stocks by a completed historical return window. Buy relative winners and short relative losers. A common 12–1 implementation skips the most recent month to reduce short-term reversal effects.

Beginner-friendlyAbout 8 minutesEvidence and practical analysis

Cross-sectional momentum: rank winners and laggardsNot a price forecast

The signal compares assets rather than forecasting one price; costs, reversals and out-of-sample results matter.

Three ways in

Understand the story. Then inspect the risk.

Start with a familiar situation, check the payoff, then test how the numbers change.

The idea · in two minutes

Rose ranks the class—but checks that every exam has been handed in

At month-end, Rose ranks stocks using a year of past results but leaves out the latest month. She buys the strongest group and shorts the weakest. Jack asks whether that means she is bullish on the whole market. No: she is asking whether winners will outperform losers.

Suppose winners gain 8% and losers fall 4% next period. A portfolio with half its capital long winners and half short losers earns 6% gross. If the market turns sharply, winners might fall 15% while losers rebound 8%; the same arithmetic then gives −11.5%. One fixed rule can have very different outcomes across regimes.

A proper study needs delisted stocks, point-in-time constituents and orders placed at the next tradable time after the signal. Momentum crashes, borrow availability, turnover, spreads and market impact can make a paper ranking expensive to trade.

趨勢延續假設Equal-weight long/short: 0.5×8%−0.5×(−4%)=6%; this is an assumed path.
市場盤整假設The relative gap narrows; turnover costs could make a small gross spread negative net.
急速反轉假設Old winners fall and old losers rebound, hurting both sides of the same portfolio.

Illustrative research examples only. They are not live signals or a recommendation.

Variants · same family, different risk levers

Thirty-nine momentum variants change the market, signal or implementation.

The list spans residual, sector, volume, cross-market and risk-managed approaches; they are not interchangeable labels for one formula.

Related strategy	Relationship	What changes	New cost or risk
Residual momentum	市場／因子中性	Regresses out market and style factors, then ranks the remaining stock-specific trend.	Residual estimates are noisy; changing factor exposures can alter the ranking.
Sector momentum rotation	產業輪動	Ranks sectors or sector ETFs by past return and rotates into stronger groups.	Sector concentration and abrupt style reversals can hit the whole portfolio.
Sector rotation with MA filter	趨勢過濾	Holds leading sectors only while their prices remain above a moving average.	The lagging filter can whipsaw in ranges and add turnover.
Dual-momentum sector rotation	相對＋絕對動量	Selects relatively strong sectors, then requires positive excess return versus cash.	Cash proxies, lookbacks and switch timing materially change exposure.
Alpha rotation	訊號組合輪替	Reweights alpha signals according to recent risk-adjusted performance.	Short samples invite performance chasing and unstable weights.
R-squared selectivity	訊號特異性	Uses return-regression R² to distinguish active selectivity from benchmark-like exposure.	R² depends on the benchmark model and is not skill or future momentum.
Regional real-estate momentum	區域／資產延伸	Rotates among regional real-estate securities using intermediate trends.	Stale valuations, leverage and regional liquidity slow the signal.
Firm-age momentum	公司生命週期	Conditions momentum on firm age or ranks stocks within age groups.	Age and survival filters create listing-history and survivorship bias.
Industry momentum	產業訊號	Uses industry returns rather than only stock returns to forecast continuation.	Common shocks can be mistaken for stock-level momentum; classifications change.
Intermediate momentum	中期持有期	Targets intermediate-horizon continuation between short-term reversal and long-run reversal.	Lookback selection can overfit and horizons differ across markets.
12-month price momentum	標準回看窗口	Ranks roughly months 12 to 2 of cumulative return, skipping the latest month.	After a market turn, old winners can become losers; momentum has crash risk.
6-month price momentum	較短回看窗口	Shortens the cumulative-return lookback to react faster to recent trends.	Reversal, noise and turnover costs become more prominent.
Junk-stock momentum	品質條件化	Studies momentum conditional on low-quality, distressed or junk-like stocks.	Illiquidity, delisting and shorting constraints can erase paper spreads.
Momentum plus long-term reversal	期限混合	Combines intermediate continuation with a long-horizon reversal adjustment.	Signal horizons and rebalance timing must be fixed before testing.
High-volume momentum	成交量條件化	Conditions price momentum on trading volume or turnover.	Volume also reflects news, attention and liquidity demand.
Fama–French residual momentum	三因子殘差	Removes market, size and value-factor returns before ranking residual trends.	Factor-window choices and residual risk make estimates unstable.
Trend factor	連續趨勢因子	Builds a factor return from trend signals across securities or assets.	The definition, scaling and shortability matter; there is no single universal formula.
52-week high	價格錨點	Measures price relative to its prior 52-week high as a continuation signal.	Splits, gaps and scale effects can distort proximity to the old high.
Six-month residual momentum	短期殘差趨勢	Uses a shorter residual-return window to capture recent stock-specific trends.	Shorter windows raise noise and trading costs.
Formulaic Alpha #1	非線性價格訊號	Uses short-horizon volatility ranks and return reversal in a nonlinear signal.	A single alpha is vulnerable to microstructure noise and universe changes.
Formulaic Alpha #52	多期限／量價訊號	Combines intermediate price position with short-term volume changes.	Complexity increases data-cleaning, parameter and cost sensitivity.
Risk-managed momentum	波動度縮放	Scales exposure using recent strategy volatility or a risk budget.	Volatility estimates lag; de-risking can coincide with a rebound.
Low-analyst-coverage momentum	資訊摩擦條件化	Tests continuation in stocks with sparse analyst coverage, where information may diffuse slowly.	Sparse coverage also means gaps, uncertainty and higher trading costs.
International momentum	跨國市場	Ranks local equity returns or relative strength across countries.	FX, non-synchronous calendars, taxes and market rules add risks.
Momentum under information uncertainty	不確定性條件化	Tests whether continuation is stronger where fundamental information is less certain.	Proxy choices change the sample; uncertainty also raises tail risk.
Corporate-bond momentum	信用債資產延伸	Ranks corporate bonds by past bond returns rather than stock returns.	Credit spreads, default, stale quotes and duration affect the measured return.
Momentum consistency and tax-loss selling	路徑一致性	Adds return-path consistency and tax-loss-selling effects to momentum.	Tax rules, calendar effects and investor composition limit portability.
Currency momentum	外匯市場	Ranks currency-pair returns, buying stronger and selling weaker currencies.	Carry, funding, hedging costs and crisis gaps mix with price momentum.
Commodity-futures momentum	商品期貨	Trades futures trends while accounting for contract rolls and term structure.	Roll yield, margin, price limits and supply shocks create distinct risks.
Momentum plus reversal in commodity futures	期貨期限混合	Combines futures trend and reversal signals across lookback horizons.	Fix horizons, contract stitching and weights; the mix may only rearrange risks.
Individualism and global momentum	國家文化交互作用	Tests whether country-level individualism covaries with international momentum.	Country traits are slow proxies and can confound institutions or development.
Momentum and mean reversion across national markets	市場層級期限比較	Compares continuation and longer-horizon reversal in national equity indices.	Few countries and common global factors reduce the effective sample size.
Market-state momentum	狀態條件化	Conditions expected momentum on the prior market state.	State boundaries can be selected after the fact and are known with a lag.
Commodity trend across horizons	期限組合	Combines commodity trend signals across investment horizons.	Horizons are correlated, and contract carry remains important.
Cognitive dissonance, sentiment and momentum	情緒機制	Links continuation to sentiment and delayed belief updating.	Sentiment measures are noisy; a mechanism result is not a tradable premium.
Momentum in FX markets	外匯參數最佳化	Studies lookback, holding-period and implementation choices for FX momentum.	Many parameter tests create data-mining risk; policy and costs change over time.
Learning to weight past returns	學習式權重	Learns weights for individual past-return observations in the signal.	More flexibility increases overfitting; strict out-of-sample tests are essential.
Liquidity-conditioned momentum	流動性條件化	Conditions momentum exposure or returns on liquidity conditions.	Spreads and funding needs worsen together in stress, so model costs may be too low.
Frog-in-the-pan momentum	資訊到達路徑	Separates gradual price adjustment from a few large jumps even at similar cumulative returns.	Requires reliable event timing; sampling density and selection can change the classification.

Fix lookback, universe, weights, rebalance and shorting rules before testing; do not assume future constituents or frictionless fills.

Read beyond the summary

References

Primary sources and research used to describe the strategy and its limits.

Price momentum
Rank assets by past returns, then buy winners and sell losers.

Understand the story. Then inspect the risk.

Rose ranks the class—but checks that every exam has been handed in

The common 12–1 signal ranks completed historical returns and skips the most recent month, which can behave differently from intermediate-horizon continuation.

Greeks: the directional intuition

Equal-weight winner-minus-loser: read the record with its caveats.

Change assumptions and inspect the scenario

Thirty-nine momentum variants change the market, signal or implementation.

References

Understand the story. Then inspect the risk.

Rose ranks the class—but checks that every exam has been handed in

The common 12–1 signal ranks completed historical returns and skips the most recent month, which can behave differently from intermediate-horizon continuation.

Greeks: the directional intuition

Equal-weight winner-minus-loser: read the record with its caveats.

Change assumptions and inspect the scenario

Thirty-nine momentum variants change the market, signal or implementation.

References

What to explore next

Earnings momentum

Lead-lag／economic links

Capital gains overhang／disposition effect

Box spread