Has Trend Gone Flat? Return Convexity in Trend Following

This paper examines return convexity in trend-following strategies, and why it appears to have declined over time. Historically, CTA and managed futures returns combined high average returns with positive convexity, a payoff that resembles being long options without the premium cost. The analysis links CTA returns to trend signals and measures how their convexity has changed, using rolling regressions of returns in rising and falling markets. It finds that the long-term signals favoured by the largest managers now show little convexity, while shorter-term and non-trend signals tend to retain more. The paper then shows how combining trend signals with cross-sectional, non-directional signals may improve the convexity and the risk-adjusted profile of a futures portfolio, and it introduces a framework for choosing among such portfolios. It is written for institutional allocators evaluating managed futures and portfolio hedges, and it frames its conclusions as analytical perspective.

What This Paper Examines

• How to measure convexity in portfolio returns relative to equity markets.
• Whether the convexity of CTA and trend-following returns has changed over time.
• Which trend signals retain convexity and which have lost it.
• How non-trend, cross-sectional signals can contribute convexity.
• How to trade off Sharpe ratio against convexity when building a futures portfolio.

Key Findings

• Convexity, not just average return, is what made trend following attractive. Convexity can be measured as the difference between a strategy’s exposure in up markets and in down markets. That separates market timing from average market exposure.
• The convexity of large CTA strategies has faded. The long-term trend signals that dominate the biggest managers show little positive convexity in recent years, particularly during equity drawdowns.
• Shorter-term trend signals tend to retain more convexity than long-term ones. Signal horizon matters, so a portfolio’s convexity depends heavily on which trend signals it emphasises.
• Non-trend, cross-sectional signals can also generate convexity. Strategies that earn more during volatile, high-dispersion periods tend to show positive convexity, even without directional market exposure.
• Combining trend and non-trend signals can improve both convexity and risk-adjusted returns. A frontier that trades Sharpe ratio against convexity helps identify portfolios that raise convexity without giving up return characteristics.

The Authors

Deepak Gurnani, Founder and Chief Investment Officer

Deepak Gurnani is the Founder and Chief Investment Officer of Versor Investments. Deepak has three decades of experience in applying quantitative methods to uncover alpha across global equity markets. Over the past decade, he has focused on pioneering the use of AI and alternative data in equity investing.

Ludger Hentschel, Founding Partner, Investment Advisor

Ludger Hentschel joined Versor Investments as a Founding Partner and is based in New York. Ludger has over 20 years of experience in quantitative research and investing.

Disclaimer: Past performance is not necessarily indicative of future results. For informational purposes only. Not an offer to sell or a solicitation of any type with respect to any securities or financial products.

Methodology: The analysis studies the SG Trend index, an average of the largest CTA hedge funds, alongside simulated trend and non-trend strategies invested in roughly 100 liquid futures contracts across equities, fixed income, commodities, and currencies. Convexity is estimated with rolling regressions that allow different market exposures in rising and falling markets, using non-overlapping monthly returns over a sample that runs from 2000 to 2022. Signals are grouped into value, momentum, and carry themes, in both time-series and cross-sectional forms. Data sources include Bloomberg and Société Générale.