Skip to content
All library documents

Allocating to Trend Following Under Uncertain Return Assumptions

Article Systematic trading blog (Rob Carver)
Read the original

Summary

This article considers how much of a portfolio to allocate to trend following when historical returns may overstate future opportunities. It compares a 60:40 US equity and bond portfolio with slow and faster trend strategies trading equity and bond futures. The author highlights weaknesses in standard optimisation, including non-normal returns, uncertain estimates, and the chance that long-term market trends will not persist. The example removes estimated returns attributable to falling bond yields and equity valuation changes, then compares correlations before and after this adjustment.

The discussion explains why trend strategies at different speeds may respond differently to a change in long-term market conditions, and why faster trend following may have a different skew profile from equities. It frames optimisation as a way to inform a practical allocation ceiling and target, while noting that the simplified three-asset example excludes commodities and other diversifying markets. The excerpt gives correlation and demeaned-return figures, but omits the optimisation results and detailed practical allocation recommendation. Its conclusions should therefore be treated as a framework for thinking about robustness, not a universal allocation rule.

Key ideas

  • Historical portfolio optimisation can be fragile when return distributions are non-normal and estimated parameters are uncertain.
  • A long period of falling yields and rising equity valuations may inflate historical estimates of future returns.
  • The example compares a 60:40 portfolio with slow and faster trend-following strategies in bond and equity futures.
  • Removing estimated secular return contributions changes the strategies’ measured correlations with the 60:40 portfolio.
  • A narrow set of markets may understate the diversification available in a broader futures trend portfolio.

Tags

Trend followingPortfolio constructionRisk managementFuturesBacktesting

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.