Dynamically Allocating Between Time-Series and Cross-Sectional Momentum
Summary
This document explains how to combine absolute, time-series momentum with relative, cross-sectional momentum in a multi-asset portfolio. It describes a unified portfolio framework that treats each pair of instruments as a relative-momentum opportunity, alongside time-series trend signals. The allocation depends on trend strength and correlations: cross-sectional momentum can gain weight when trends differ across instruments and their returns are sufficiently correlated.
The article summarizes a test of four strategies using 59 equity-index, interest-rate, currency, and commodity futures from 2001 through 2018, with three moving-average parameter sets. It reports that the dynamically weighted strategy had stronger Sharpe performance and lower turnover than pure strategies or a fixed 50/50 blend, including after transaction costs. The authors estimate that time-series momentum supplied about 80% of the dynamic portfolio's allocation on average. These findings are historical backtest results; the supplied text omits the underlying equations, detailed tables, and full implementation details, so it does not establish that the results will persist in other markets or periods.
Key ideas
- Time-series momentum takes directional positions based on an instrument’s own trend, while cross-sectional momentum ranks instruments against one another.
- The proposed framework represents every instrument pair as a relative-momentum opportunity.
- The balance between absolute and relative momentum depends on trend strength and cross-instrument correlation.
- In the reported historical tests, dynamic allocation outperformed a fixed equal-weight blend, while time-series momentum contributed most of the allocation on average.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.