The interconnections between U.S. financial CDS spreads and control variables: New evidence using partial and multivariate wavelet coherences

Abstract

We examine the connectedness of the sector CDS spreads for U.S. banks, financial services firms and insurers with major global factors including the US stock market volatility, Libor, Treasury bill rates and oil prices to gauge credit risk exposure. Our main objective is to gain insight into the co-movement between those CDSs and WTI and the other control variables, which reflect the health of the economy. The partial wavelet and the multi-wavelet coherence are applied to the available CDS data. The findings show strong evidence of partial wavelet coherence at the medium and low frequencies (time horizons) between the three sector CDS, and that there is a strong short-term co-movements between the CDS spreads and their control variables at high frequencies (short horizons). The multivariate wavelet analysis shows that the three CDS spreads are generally highly correlated over time and frequency bands. In addition, we uncover some differences between the co-movements of the three CDSs with each other and in their relationships with the control variables through time and frequency. Furthermore, the strength of the partial and multiple wavelet co-movements varies through time and frequency and does not subside after the Great Recession ended and during the ensuing slow recovery that started in the second half of 2009. These results are useful for forecasting sector CDS spreads and establishing their interrelationship with global US financial and oil markets’ factors.