基于复杂网络的汇率市场关联网络结构实证研究

发布时间：2018-03-19 19:07

本文选题：复杂网络　切入点：汇率　出处：《广西大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着世界经济一体化的推进,国际竞争与资本流动日趋激烈、频繁,金融市场不确定性多,呈现复杂的波动趋势。汇率市场作为重要的金融市场,影响国际收支,对于汇率波动规律的研究,有助于帮助货币当局及时应对汇率市场的波动变化。由于汇率受到利率、通货膨胀等诸多因素影响,汇率波动发生及其演化行为表现出复杂系统的特点,传统国际金融理论、汇率决定理论对于汇率市场波动性的解释越来越难。汇率市场网络连接复杂,各个国家构成的货币节点丰富,以汇率波动相关性形成的汇率网络波动变化复杂,具有典型的复杂网络形式,分析汇率市场网络拥有何种拓扑性质及变化规律,为研究汇率市场波动情况提供了新的途径。学界一直将复杂网络的研究方法应用在股票市场与股票指数上,对于汇率市场构成的关联网络的研究还在起步阶段,相关研究成果少。本文根据复杂网络理论,利用整体网分析方法,选取2007年1月到2016年12月期间一共3128个交易日以美元标价的52种货币数据,计算货币间相关系数,选取各年度相关系数均值作为建立汇率关联网络的阈值,采用阈值法构建各年度全连无向无权网络。分析汇率网络的拓扑性质,探究汇率市场波动的规律。本文通过UCINET6网络分析软件计算相关指标,发现汇率市场关联网络具有较短的平均路径长度,十年间最短路径长度是1和2时包含了几乎90%的情形,聚类系数高,徘徊在0.8上下,以此验证汇率网络存在小世界效应。分析汇率网络中各节点度中心性,对货币进行权力性分析,对汇率网络内的节点进行重要性排序,最后,针对推进人民币国际化进程提出相关建议。
[Abstract]:With the development of world economic integration, international competition and capital flow are becoming more and more intense and frequent, and the financial market is uncertain, showing complex fluctuation trend. As an important financial market, the exchange rate market affects the balance of payments. The study of the law of exchange rate fluctuation is helpful to help monetary authorities cope with the fluctuation of exchange rate market in time. Because the exchange rate is affected by many factors such as interest rate, inflation and so on, The occurrence and evolution of exchange rate fluctuations show the characteristics of complex systems. The traditional international financial theory and exchange rate determination theory are more and more difficult to explain the volatility of exchange rate market. Each country has abundant currency nodes, and the exchange rate network formed by the correlation of exchange rate fluctuations is complex and has a typical complex network form. This paper analyzes the topological nature and variation law of the exchange rate market network. It provides a new way to study the fluctuation of exchange rate market. The research method of complex network has been applied to stock market and stock index in academic circles, and the research on the related network of exchange rate market is still in its infancy. According to the theory of complex network and using the method of whole network analysis, this paper selects 52 kinds of currency data, which are marked in US dollars for a total of 3128 trading days from January 2007 to December 2016, and calculates the correlation coefficient between currencies. This paper selects the annual correlation coefficient mean as the threshold to establish the exchange rate correlation network, and uses the threshold method to construct the undirected undirected network in each year. The topological properties of the exchange rate network are analyzed. In this paper, it is found that the correlation network of exchange rate market has a short average path length, and the shortest path length in ten years is 1 and 2:00, which includes almost 90%, through the calculation of related indexes by UCINET6 network analysis software. The clustering coefficient is high, hovering around 0.8, which verifies the existence of small-world effect in the exchange rate network. This paper analyzes the centrality of each node in the exchange rate network, analyzes the currency power, and sorts the importance of the nodes in the exchange rate network. To promote the internationalization of the RMB process to put forward relevant recommendations.
【学位授予单位】：广西大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.6

【参考文献】