Time Series Momentum
Read the paperopens doi.org in a new tab
What they found
Unlike cross-sectional momentum (buy the winners relative to other assets), time-series momentum asks whether each asset's own past return predicts its own future return. Across 58 liquid futures and forwards in equities, bonds, currencies, and commodities from 1985 to 2009, an asset's return over the past 12 months positively predicted its return over the next month, and a simple strategy of going long assets with positive 12-month returns and short those with negative ones produced strong, diversified returns. Returns were partly reversed after about a year, and the strategy did best in extreme market moves, resembling a long straddle.
See it drawn
Original diagrams for the ideas on this page. Illustrative, not real market data.
What you can use
- The simplest trend rule, 'is the 12-month return positive or negative', works across nearly every liquid futures market.
- Trend following historically made money in the worst equity months, which is why it is used as crisis diversification.
- Scale each position by its volatility so that no single market dominates the book.
- Trend signals decay after roughly a year, so refresh them; do not cling to a position because it once trended.
Caveats
Managed-futures-style portfolio with dozens of markets; a single-market version is far noisier. Gross of costs, though the markets are liquid. Trend following had a long flat stretch from 2009 to 2019 that this sample largely predates.
Tags: momentum, trend-following, futures, multi-asset
Summaries are our own reading of the paper, not the authors' words. Educational only, not advice. Discuss it in Book Club.