Oil And US Dollar Exchange Rate Long-run Cross-correlation

This paper studies the long-run cross-correlations and their asymmetric behavior between oil prices and US dollar exchange rates by employing the detrended cross-correlation analysis,multifractal detrended cross-correlation analysis and their noval asymmetric extension methods.First,we investigate the long-run cross-correlations between oil prices and exchange rates in a global scope.The DCCA coefficient can reveal the strength of dependence in different scales.The results imply that,oil prices and other bilateral exchange rates are positively correlated except Japanese Yen.When oil price surges,USD and Yen depreciate while other currencies appreciate.Dynamic tests reveal that,the cross-correlations are relatively low before financial crisis but become significant in the aftermath of financial crisis.The results of MF-DCCA provide evidence that their cross-correlations are multifractal.They are persistent when facing small fluctuations but anti-persistent when facing large fluctuations.Then we separate the different trends of oil prices to characterize the asymmetries hidden in the cross-correlations.The results of ADCCA imply that,the dependence is comparatively weak when oil price is in upward trend while strong when it is in downward trend.We then use the traditional econometric models including "mean spillover" and "volatility spillover" to filter each pair of oil price and exchange rate.But it turns out that,asymmetric cross-correlations between filtered pairs still exist significantly.We also find that,the past trends of oil prices appear to affect the dependence structures symmetrically.Finally,MF-ADCCA results indicate persistence in cross-correlations becomes stronger when oil price is in upward trend.