Abstract
This study first decomposes the daily returns of regional gas and crude oil at different time scales, using the ensemble empirical mode decomposition (EEMD) method. It then investigates the causality relationship between each pair of components at the different time scales, by employing the linear and nonlinear Granger causality tests. For the original returns series, this study finds that unidirectional linear Granger causality exists from crude oil markets to North American and European gas markets. However, for nonlinear characteristics, the crude oil and regional gas markets exhibit bidirectional nonlinear Granger causality. On the medium-term time scale, a bidirectional nonlinear spillover effect is found between the markets. The long-term trends for the markets suggest a significant linear relationship; however, no nonlinear spillover effect is found between the crude oil and regional gas markets.
| Original language | English |
|---|---|
| Pages (from-to) | 98-110 |
| Number of pages | 13 |
| Journal | Energy Economics |
| Volume | 67 |
| DOIs | |
| State | Published - Sep 2017 |
Keywords
- Crude oil market
- Ensemble empirical mode decomposition
- Multi-scale analysis
- Nonlinear Granger causality test
- Regional natural gas market
Fingerprint
Dive into the research topics of 'The relationship between regional natural gas markets and crude oil markets from a multi-scale nonlinear Granger causality perspective'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver