Skew–Brownian processes for estimating the volatility of crude oil Brent

Abstract

To predict the volatility of crude oil Brent price, we propose a novel econometric model ¹ where the explanatory variables are a combination of macroeconomic variables (i.e. price pressure), trade data (freight shipment index), and market sentiment (gold volatility). The model is proposed in two alternative variants: first, we assume Gaussian distributed quantities; alternatively, we consider the potential presence of skewness and adopt a Skew–Brownian process. We show that the suggested approach outperforms the selected baseline model as well as other models proposed in the literature, especially when turbulent periods occur.