全球外汇市场网络结构、货币影响力与货币社区

Topology of Foreign Exchange Markets,Currency Influence Strength and Community Detection

文章采用小波变换法和平面极大滤图法,第一次从多频视角分析了由全球45种主要货币组成的外汇市场的网络结构,并进一步研究了相关货币的影响力及货币社区。研究发现:货币的汇率波动在低频周期上体现了更高的相关性。外汇市场网络呈现"无标度"和"小世界"两个复杂网络特征,并在高频周期上显现了更高的内聚性。不同频率下的外汇市场根据货币间的相依结构形成了4个或5个社区集团。货币社区的分区在高频周期上主要体现了地域性,在低频周期上则主要体现了国家间的深度经济联系。

In this study,we employ a wavelet-based approach and the Planar Maximally Filtered Graph( PMFG) method to study the topology of 45 major currencies in the foreign exchange( FX) markets from a view of multi-resolution decomposition. Based on the network of FX markets,we then derive a strength index of currency influence and detect the currency communities. The resultshows that the correlation between currencies on a low-frequency scale is higher than the correlation on a high-frequency scale.The PMFG of FX markets is a scale free network with the small world property and it has higher clustering coefficient on a highfrequency scale. We also provide the top 10 and bottom 10 currencies of influence strength on four scales. Finally,this paper find 4 to 5 currency communities in the FX markets on four scales. The regionalism has a great effect on community division on a highfrequency scale and the deep economic ties between countries has a great effect on community division on a low-frequency scale.