摘要
从汇率行为的非线性动态复杂性出发,从理论上探讨了汇率系统非线性依赖关系对描述和理解汇率行为的重要性,结合相空间重构和替代数据两类具体的非线性系统分析方法,利用多种统计检验工具,依次对四个不同时间跨度区间上人民币兑美元汇率时间序列数据中是否存在非线性依赖关系的统计显著性进行了检验。实证结果表明,人民币汇率系统是一个典型的非线性动态复杂系统,不同时间跨度上的人民币汇率收益率序列均为包含了某种非线性依赖关系的时间序列数据,并且这些非线性依赖结构均表现出了一定的高维特征,此外人民币汇率系统内部各要素之间的可检验非线性依赖关系在时间上也表现出了一定的趋同性。
Based on the nonlinear dynamic complexity of exchange rate behavior, this paper discusses the relationship between nonlinear dependence embedded in exchange system and it's behavior theoretically. Then under the phase reconstruction and surrogate data analyzing framework, we aim at four different timing ,scale RMB/USD exchange rate time series, and test if there is apparent statistically nonlinear dependence embedded in respective series. The empirical results confirm the positive conclusion and indicate specifically that the RMB exchange rate yield series are typically nonlinear structure data and all the nonlinear dependency structure appear high dimensional characteristics and some show nonlinear convergence between the internal factors in the studied exchange rate system.
出处
《系统工程》
CSSCI
CSCD
北大核心
2016年第6期43-49,共7页
Systems Engineering
基金
国家自然科学基金资助项目(71373072
71340014)
高等学校博士点专项科研基金资助项目(20130161110031)