Abstract: In recent years, the price trend among international oil markets has shown inconsistency, which raises doubts about the traditional assertion of the consistency of inter-market linkages.With the help of wavelet transform and cross-spectral analysis, this paper chooses WTI-LLS-Brent daily price data of three international oil markets from 2006 to 2016 for correlation study.Empirical results show that market correlation is inconsistent in different frequency perspectives:the correlation of each high frequency component increases with the decrease of frequency, and the Granger causality test results of the low frequency trend series and the original price series are quite different;the information contained in the high frequency signal helps to predict the short-term trend of the market, while the long-term trend of the oil market price is difficult to predict.Research shows that financial factors have short-term impact on high-frequency linkage, while changes in supply and demand have far-reaching impact on market linkage, while policy changes have no obvious impact on market linkage. Because of the different nature of the market and location factors, different market responses to different types of shocks have resulted in inconsistent correlations.The study suggests that different market regulation measures should be adopted on the basis of identifying the relevant factors. At the same time, we should promote the construction of spot market and pay attention to improving the oil market system so as to diversify the means of market risk management to withstand external shocks.

Key words: risk management, co-movement, wavelet transform, Copula, Wavelet coherency