Is Nonlinear Momentum the Answer?

Key Takeaways

• Nonlinear momentum dampens extreme returns, aiming to curb crashes.
• Study shows higher Sharpe ratios vs linear model in many periods.
• Recent decade sees nonlinear approach underperforming traditional momentum.
• Performance gap narrows in bear markets, suggesting environment sensitivity.
• Robustness concerns limit adoption despite theoretical advantages.

Pulse Analysis

Momentum investing remains a cornerstone of quantitative portfolios, yet its linear formulation—simply ranking assets by past 12‑month returns—exposes portfolios to sharp reversals when extreme winners tumble. The recent work by Moskowitz and collaborators proposes a nonlinear transformation that compresses outsized gains, theoretically smoothing the return distribution and lowering crash risk. Early back‑tests reveal higher Sharpe ratios across multiple horizons, particularly when market conditions turn bearish, indicating that the model may capture hidden risk premia missed by the classic approach.

Despite these promising signals, the last decade tells a different story. The nonlinear strategy’s edge has eroded, and in some periods it trails the conventional momentum benchmark. Analysts point to a broader slowdown in momentum profitability, driven by higher transaction costs, faster information diffusion, and increased competition among algorithmic traders. In such an environment, the added complexity of nonlinear weighting may not translate into superior net returns, especially after accounting for implementation frictions.

For practitioners, the key takeaway is a nuanced risk‑return trade‑off. While nonlinear momentum can enhance performance during turbulent phases, its long‑term robustness remains unproven. Investors must weigh the potential upside against higher model risk, data requirements, and operational costs. As markets continue to evolve, a hybrid approach—leveraging both linear and nonlinear signals—might offer a pragmatic path forward, balancing simplicity with the desire for crash mitigation.