基于多目标决策与数据挖掘融合方法的主权信用风险评估研究

发布时间：2019-03-23 19:24

【摘要】：2007年初美国爆发了本世纪以来影响最大的次级抵押贷款危机,危机迅速蔓延至欧洲大陆诸国,而早期预警系统、信用评估体系、信用评级机构却在此次危机面前集体失效。金融创新使得全球化投资主体无法小视任何金融风险,特别是结构化的金融产品。全球交易系统已将危机传递给了全世界,涵盖了金融体系内的所有机构。从发生的地域来看,此次危机从美国向欧洲、日本乃至全球迅速扩散,呈现了全球性和系统性的特征。传统经济学理论认为：金融危机更容易在监管不完备、金融市场与本国经济结构具有明显错配的新兴市场国家爆发,而此次危机却源自金融业市场化程度最高的美国。由此可见,复杂的结构化金融产品和针对金融监管的创新,对现代复杂金融环境下的危机预警提出了更高的要求。目前尚不存在对创新引发金融危机有效的预警方法,因此需要我们改变研究思路,寻找新方法解决这个古老问题。本研究尝试引入新兴交叉学科的成果,提高防范创新型金融危机爆发后带来的国家主权信用违约风险的能力。鉴于次级抵押贷款危机所表现出来的复杂性和全球化特点,我们对国家主权风险的研究内容界定为：国家主权信用违约风险的数量化排序问题。研究思路是借助多目标决策方法,为创新型金融危机后的国家主权信用违约风险排序提供方法支持；借助数据挖掘技术,对量化排序的国家主权信用违约风险进行分级评价。本研究的目标是：探索解决新型金融危机全球化带来的国家主权信用违约风险识别与防范问题,为全球范围内的贸易与投资,提供国家主权信用违约风险研判与规避的科学方法。基于上述考虑,本研究论文包括以下内容：第一、回顾了次级抵押贷款危机逐步发展成为金融危机的全过程,并指出了基于全球化金融创新以及现代复杂金融产品所带来的创新型金融以及由此引发的新问题,提出了引入新兴交叉学科的研究方法以应对新出现问题的研究思路。第二、次级抵押贷款危机被认为是一场信用风险所引发的金融危机。文中回顾了已有的信用风险评估模型与信用评级等相关概念,以次级抵押贷款危机逐步发展成为金融危机并导致国家主权信用违约为线索,指出目前已有的国家主权信用评级方法存在的不足,利用计量经济学的传统方法预测了由次贷危机所导致国家主权信用违约风险,得出结论：传统计量经济学方法无法预测、识别这类由金融创新导致的国家主权信用风险。第三、数据挖掘与多目标决策方法的发展为解决国家主权信用违约问题提出了新的思路和数量化方法。研究中利用数据挖掘与多种多目标决策方法的融合,结合次级抵押贷款危机的特点,提出了基于时间维度的多目标风险预警模型,对国家主权信用违约风险的量化排序进行了实证研究。第四、研究的最后,对国家主权信用违约风险的量化排序结果利用数据挖掘方法进行分级评价,并借鉴金融压力测试的思想,对模型进行了敏感性分析和进一步的结果检验。通过实证检验证明了本文提出的国家主权信用违约风险评估模型具有良好的预测效果,通过模型检验和敏感性分析,证明了模型具有良好的鲁棒性和稳定性。国家主权信用违约风险是一个涉及宏观与微观的多层面的研究对象,研究内容宽泛,数据类型多样,属于复杂系统的研究领域,已有的研究中尚未形成被广泛接受的统一研究范式。本研究作为该领域的探索性研究之一,尝试将新兴交叉学科的研究方法引入其中,为古老的国家主权信用违约风险问题提供新的研究思路。
[Abstract]:In early 2007, the U. S. outbreak of the largest subprime mortgage crisis since the century, the crisis spread rapidly to the countries of the European continent, and the early warning system, the credit evaluation system and the credit rating agencies were collectively failing in the face of the crisis. The financial innovation makes the main body of the globalization not to see any financial risks, especially the structured financial products. The global trading system has delivered the crisis to the world, covering all the institutions in the financial system. From the regional perspective, the crisis has spread rapidly from the United States to Europe, Japan and the world, presenting global and systematic features. The traditional economics theory holds that the financial crisis is more easily regulated and the financial market has a clearly mismatched emerging market country with its own economic structure, and the crisis has come from the most market-oriented United States in the financial industry. It can be seen that the complex structured financial products and the innovation of financial supervision have put forward higher requirements for the crisis early warning in the modern complex financial environment. There is no effective early-warning method for innovation-induced financial crisis, so we need to change the research thinking and find a new way to solve this old problem. The study attempts to introduce the results of the emerging cross-discipline and improve the ability of the risk of national sovereignty credit default after the outbreak of the innovative financial crisis. In view of the complexity of the subprime mortgage crisis and the characteristics of globalization, our study on the risk of national sovereignty is defined as the quantitative sequencing of the risk of national sovereignty credit default. In this paper, by means of the multi-objective decision-making method, the paper provides the method support for the national sovereignty credit default risk sorting after the innovative financial crisis; by means of the data mining technology, the national sovereignty credit default risk of the quantitative ordering is graded and evaluated. The purpose of this study is to explore the national sovereignty credit default risk identification and prevention problems arising from the globalization of the new financial crisis, to provide the scientific side of the research and avoidance of the national sovereignty credit default risk for trade and investment in the global scope Based on the above considerations, the present study includes the following First, it reviews the step-by-step development of the subprime mortgage crisis as the financial crisis The whole process, and points out the innovative finance brought about by the financial innovation of globalization and the modern complex financial products, and the new problems arising from it, put forward the research method to introduce the new cross-discipline to deal with the new problems. The second, the subprime mortgage crisis is considered a credit risk. In this paper, the related concepts of credit risk assessment model and credit rating are reviewed. The development of the secondary mortgage crisis has become the financial crisis and leads to the sovereign credit default of the national sovereignty, and points out the existing method of national sovereignty credit rating. The traditional method of econometrics has been used to predict the risk of the sovereign credit default caused by the sub-loan crisis, and it is concluded that the traditional econometrics method is not able to predict and identify the state-owned enterprises caused by the financial innovation. Third, the development of data mining and multi-objective decision-making is a new way to solve the problem of sovereign credit default. In this paper, a multi-objective risk early-warning model based on time dimension is put forward based on the combination of data mining and multi-objective decision-making methods, and the quantitative ordering of the risk of sovereign credit default is presented. The fourth, the last of the research, the quantitative ranking result of the national sovereignty credit default risk is classified and evaluated by the data mining method, and the model is sensitive by using the thought of the financial pressure test. inductive analysis and entry The results of one-step results show that the model of national sovereignty credit default risk assessment presented in this paper has a good prediction effect, and it is proved that the model has a good model by means of model test and sensitivity analysis. Robustness and stability. The risk of national sovereignty credit default is a multi-level research object involving macro and micro, the content of the study is broad and the data type is diverse, which belongs to the research field of complex system, and has not been widely accepted in the existing research. This study, as one of the exploratory studies in this field, has tried to introduce the research methods of emerging cross-disciplines into one of them, and to raise the risk of default of the ancient national sovereignty
【学位授予单位】：电子科技大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：F224;F831

【参考文献】