Wavelet Multiscale and Spillover Analyses of Volatility and Correlation

Abstract

This article investigates the dynamic empirical relationships among realized, risk-neutral, and risk premium measures of volatility and correlation of the S&P 500 stock index from January 1, 2000, to December 31, 2020. The empirical investigation runs a spillover analysis to identify the receiver and the transmitter variables and implements a wavelet local multiple correlation (WLMC) methodology to study the multiscale correlations. The results identify the implied measures as the most influential variables and also reveal that the strength of correlation is changing with time scales; moreover, the correlation between volatility risk premium and correlation risk premium is not always statistically significant through either time scales or time periods. These findings support the use of scale-based correlation metrics in derivatives studies.