论文部分内容阅读
国家风险是经济活动主体在国际业务中所面临的来自其他国家的风险,深入研究其内在特征,对于理解和把握国家风险的动态演化规律有着重要意义。鉴于国家风险复杂易变的特点,本文提出了一种基于“分解重构”思想的多尺度特征提取与识别的研究框架,利用Ensemble EMD方法将原始国家风险值分解到短期、中期和长期三个时间尺度上,引入方差贡献率、相关系数和Shapley值刻画各尺度与原始国家风险序列间的波动特征、模态特征以及全局重要度。以12个OPEC石油输出国为样本,实证结果发现:利用各尺度的模态特征和波动特征可以很好地实现样本国国家风险的分类管理,且分类具有较好的一致性;由Shapley值获得不同尺度的全局重要度,对于全部样本国呈现出一致且稳定的内在固有特征,即短期、中期和长期三尺度对国家风险的“贡献度”约为1∶1∶3。这不仅能够为国家风险管理提供了更为丰富的动态特征信息,而且对于更为全面的国家风险特征识别、监测与预测提供了一种新的研究方法。
Country risk is the risk that the main body of economic activity faces from other countries in its international business. It is of great significance to study its inherent characteristics in depth to understand and grasp the dynamic evolution of national risk. In view of the characteristics of complex and variable national risks, this paper presents a research framework for multi-scale feature extraction and recognition based on the idea of “decomposition and reconstruction ”, using the Ensemble EMD method to decompose original country risk values into short, medium and long term On three time scales, the variance contribution rate, correlation coefficient and Shapley value are introduced to characterize the fluctuation characteristics, modal characteristics and global importance of all the risk sequences in the original country. Taking 12 OPEC countries as samples, the empirical results show that the classification and management of the national risk in the sample countries can be well achieved by using the modal characteristics and the volatility characteristics of each scale, and the classification has good consistency. The value obtained from the Shapley value The global importance of different scales shows a consistent and stable inherent characteristic for all sample countries, that is, the “contribution” of short-term, medium-term and long-term risk to country risk is about 1: 1: 3. This not only provides richer dynamic feature information for national risk management, but also provides a new research method for more comprehensive identification, monitoring and prediction of national risk characteristics.