Cross-correlation between crude oil and refined product prices

In this paper, we investigate cross-correlations between crude oil and refined product prices based on the well-known detrended cross-correlation analysis (DCCA). Our findings indicate that the cross-correlations are significant and strong. Furthermore, the multifractality in cross-correlations is also revealed. The cross-correlation coefficients are as high as 0.9 for larger time scales and are greater than those for smaller time scales. Two popular models, vector error correction model and bivariate BEKK volatility model, are found to have very limited power in capturing long-range cross-correlations, suggesting the drawbacks of these conventional models in actual applications. Long-term cross-correlations are stronger in recent ten years than those in the past decades.