Allocating Trend-Following Strategies Across Correlated Assets

Summary

This study considers how to allocate trend-following strategies across multiple assets whose returns may be correlated. Under Gaussian-market assumptions and linear strategies, it derives formulas for portfolio mean and variance, then uses them to form a risk-adjusted allocation. The authors show that a dynamic allocation across n assets can be represented as a static allocation over n-squared virtual assets, with lead-lag adjustments to trend-following positions.

The paper investigates asset autocorrelation and cross-asset correlation in a two-asset setting and a sector model. It argues that correlations can help estimate apparent trends and adjust positions, so they may improve allocation rather than simply reduce diversification benefits. The conclusions depend on the simplifying Gaussian and linear assumptions, and the excerpt provides no empirical performance results or implementation details. Its main contribution is an analytical framework for thinking about how correlations interact with trend-following signals and portfolio construction.

Key ideas

• The paper derives risk-adjusted allocations for linear trend-following strategies under Gaussian assumptions.
• A dynamic allocation problem across n assets is reformulated using n-squared virtual assets and lead-lag adjustments.
• Both asset autocorrelation and cross-asset correlation affect trend-following portfolio behavior.
• The authors argue that cross-asset correlations can improve trend estimation and position adjustment.
• The excerpt provides a theoretical framework but no reported empirical performance evidence.

Tags

# Optimal Allocation of Trend Following Strategies


# Optimal Allocation of Trend Following Strategies









We consider a portfolio allocation problem for trend following (TF) strategies on multiple correlated assets. Under simplifying assumptions of a Gaussian market and linear TF strategies, we derive analytical formulas for the mean and variance of the portfolio return. We construct then the optimal portfolio that maximizes risk-adjusted return by accounting for inter-asset correlations. The dynamic allocation problem for $n$ assets is shown to be equivalent to the classical static allocation problem for $n^2$ virtual assets that include lead-lag corrections in positions of TF strategies. The respective roles of asset auto-correlations and inter-asset correlations are investigated in depth for the two-asset case and a sector model. In contrast to the principle of diversification suggesting to treat uncorrelated assets, we show that inter-asset correlations allow one to estimate apparent trends more reliably and to adjust the TF positions more efficiently. If properly accounted for, inter-asset correlations are not deteriorative but beneficial for portfolio management that can open new profit opportunities for trend followers.

Shown in full with attribution under the source's licence. Licence: abstract CC0

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