基于互信息熵的国家风险相关性研究  

Correlation research of country risk based on mutual information

作　　者：姚晓阳[1,2] 孙晓蕾[1] 吴登生[1] 杨玉英[1,2] 

机构地区：[1]中国科学院科技政策与管理科学研究所,北京100190 [2]中国科学院大学,北京100049

出　　处：《系统工程理论与实践》2015年第7期1657-1665,共9页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(71003091;71373009;71133005);中国科学院青年创新促进会项目

摘　　要：国家间国家风险的相互关联状况,已成为影响全球投资战略和国际资本流动的重要决定因素,因此有必要深入识别国家风险相关性特征.本文基于信息熵理论,引入互信息构建广义相关系数刻画国家风险间复杂的相关关系,并通过构建"多阶段一多要素"的分析框架,刻画了2007年金融危机前后政治、经济、金融三种国家风险要素的相关性特征差异.以金砖国家为例,研究发现:金融危机前后综合国家风险相关性变化明显;即使国家间综合国家风险相关性变化趋势相似,但其内在的国家风险要素相关性特征差异显著.研究结果有利于国际投资者规避来自国家层面的不确定性损失,对国际贸易与投资有着重要的指导意义.Correlation of country risk between different countries has been a crucial influencing factor for worldwide investment strategy and international capital flow. Given this, identifying the correlation characteristics is of great importance. In this paper, mutual information is introduced based on the information entropy theory to construct globalized correlation, which will be used to depict the complex correlation relationship between different countries＇ country risk. Further, a ‘multi-phase ＆ multi-elements’ frame is proposed and correlation characteristic differences before and after financial crisis in 2007 for political/economic/financial risk are given respectively. BRICS countries are selected as the sample countries, and results as follows are obtained. For composite country risk, after the crisis correlation between the five countries changes obviously compared with its behavior before the financial crisis. Even if correlation of composite country risk changes in the same way, the correlation characteristics for different country risk elements may also behave differently. Results obtained in this paper can not only help international investors avoiding lost coming from country risk, but also offer guidance for international trade and investment.

关 键 词：风险相关性 互信息熵 国家风险 广义相关系数 

分 类 号：F224[经济管理—国民经济]