Linear and nonlinear Granger causality investigation between carbon market and crude oil market: A multi-scale approach
Linear and nonlinear Granger causality investigation between carbon market and crude oil market: A multi-scale approach
复制标题
碳市场与原油市场之间的线性和非线性格兰杰因果关系研究：多尺度方法
DOI：
10.1016/j.eneco.2015.07.005
复制
发表时间：
2015-09
期刊：
影响因子：
12.8
通讯作者：
Wang, Shuai
中科院分区：
文献类型：
--
作者：
Yu, Lean;Li, Jingjing;Tang, Ling;Wang, Shuai
本文采用多尺度分析方法对碳市场与原油市场之间的因果关系进行了研究，主要包括两个步骤：多尺度分析和因果关系检验。在多尺度分析中，采用二元经验模态分解（BEMD）对两个不同时间尺度的市场收益序列进行分解。在因果检验中，建立了线性和非线性的综合格兰杰检验来研究在相似的时间尺度上每对匹配分量之间的关系。以欧盟排放限额（EUA）期货和布伦特（Brent）期货为研究样本，可以得到一些有趣的发现。(1)在原始数据层面（未进行多尺度分解），本研究发现了支持中性假设的证据，即碳与原油市场之间不存在格兰杰因果关系。(2)在小时间尺度上（一周内不包括非工作日），两个市场可能是不相关的，并由各自的供需不平衡驱动。(3)在中等时间尺度（一周以上一年以下）下，由于某些具有中期效应的额外因素，如重大事件和政策变化，两个市场之间存在较强的双向线性和非线性溢出效应。(4)在长时间尺度上，两个市场的长期趋势呈现出明显的线性关系。
This paper investigates the causality between carbon market and crude oil market using a multi-scale analysis approach, in which two main steps are involved: multi-scale analysis and causality testing. In multi-scale analysis, bivariate empirical mode decomposition (BEMD) is employed to decompose the two series of market returns at different time-scales. In causality testing, a linear and nonlinear integrated Granger test is formulated to investigate the relationship among each pair of matched components on a similar time-scale. With the European Union emission allowance (EUA) futures and Brent futures as study samples, some interesting findings can be obtained. (1) At the original data level (without multi-scale decomposition), this study finds evidence supporting a neutrality hypothesis, i.e., no Granger causality between the carbon and crude oil markets. (2) On small time-scale (within one week excluding non-work days), the two markets might be uncorrelated and driven by their own respective supply–demand disequilibriums. (3) For medium time-scale (above one week but below one year), there is a strong bi-directional linear and nonlinear spillover effect between the two markets, due to certain extra factors with medium-term effects, e.g., significant events and policy changes. (4) For long time-scale, the long-term trends of the two markets appear an obvious linear relationship.
登录
查看更多内容
基于压缩感知的原油价格预测人工智能学习范式
影响因子：
12.8
作者：
Yu, Lean;Zhao, Yang;Tang, Ling
通讯作者：
Tang, Ling
影响因子：
3.9
作者：
Tanaka, Toshihisa;Mandic, Danilo P.
通讯作者：
Mandic, Danilo P.
影响因子：
4.8
作者：
L. Tang;Shuai Wang;Kaijian He;Shouyang Wang
通讯作者：
L. Tang;Shuai Wang;Kaijian He;Shouyang Wang
DOI：
10.1016/j.matcom.2014.12.001
发表时间：
2015-05
期刊：
Math. Comput. Simul.
影响因子：
--
作者：
J. Reboredo;Mikel Ugando
通讯作者：
J. Reboredo;Mikel Ugando
影响因子：
4.7
作者：
J. Reboredo
通讯作者：
J. Reboredo