Dynamic spillover effects among international crude oil markets from the time-frequency perspective

This research takes seven representative crude oil markets in the world, decomposes and reconstructs the yield series by CEEMDAN and Fine-to-coarse algorithm, and measures the markets’ risk level by applying the DCC-GARCH-CoVaR model. We further construct a network spillover model based on TVP-VAR to investigate the return spillover and risk spillover effects among these oil markets in different time scales. The empirical results are as follows. (1) Integration within the international crude oil market is deepening, and return spillover and risk spillover are at high levels. (2) In the short run, Brent, Tapis, and Bonny crude oil markets are the main net exporters of return spillovers, while in the long run, Brent, WTI, and Dubai crude oil markets are global crude oil price benchmarks. (3) The risk level of each crude oil market under the full sample and high-frequency perspective is generally consistent, and the dynamic spillover effects between markets are relatively close, while the Brent and Tapis crude oil markets are the main net exporters of risk spillovers from the low-frequency perspective. (4) The impact of the same event on the spillover effect is heterogeneous in different time scales. For example, the spread of the COVID-19 epidemic in 2020 and the break-up of the “OPEC +” crude oil negotiations reduced the risk spillover level in the short term, but increased the risk spillover level in the long term.