Dynamic Multiscale Relationships Between COVID-19 Sentiment and Extreme Crude Oil Returns: Evidence from Wavelet Coherence Analysis

Abstract

ABSTRACT This study investigates the dynamic, multi-scale relationship between sentiment related to the COVID-19 pandemic and extreme returns in crude oil. The recently developed COVID-19 indices are employed to gauge pandemic sentiment. Utilizing daily data spanning from January 2020 to December 2021, Granger’s linear and nonlinear causality tests reveal that indices nonlinearly influence extreme fluctuations in West Texas Intermediate and Brent crude oil prices. Interestingly, a reciprocal causation is also identified: extreme crude oil returns significantly affect the indices. Furthermore, the wavelet transform coherence analysis sheds light on the indices’ ability to predict extreme crude oil price volatility across specific time-frequency domains, displaying diverse distributions and lead-lag patterns among the sub-indices. Our study underscores the efficacy of indices in anticipating extreme fluctuations in crude oil values during the COVID-19 pandemic, carrying important implications for investors, scholars, and policymakers.