Implied volatility surface predictability: the case of commodity markets

TL;DR

This paper examines whether existing models can predict implied volatility surfaces in commodity markets, finding that modeling the term structure with Nelson-Siegel factors yields the most accurate forecasts for energy and precious metals options.

Contribution

It demonstrates the effectiveness of Nelson-Siegel factor models in forecasting implied volatility surfaces in commodity markets, filling a gap in existing literature.

Findings

Nelson-Siegel factors improve forecast accuracy.

Energy and precious metals options show predictable volatility patterns.

Explicit term structure modeling outperforms other approaches.

Abstract

Recent literature seek to forecast implied volatility derived from equity, index, foreign exchange, and interest rate options using latent factor and parametric frameworks. Motivated by increased public attention borne out of the financialization of futures markets in the early 2000s, we investigate if these extant models can uncover predictable patterns in the implied volatility surfaces of the most actively traded commodity options between 2006 and 2016. Adopting a rolling out-of-sample forecasting framework that addresses the common multiple comparisons problem, we establish that, for energy and precious metals options, explicitly modeling the term structure of implied volatility using the Nelson-Siegel factors produces the most accurate forecasts.