【作者】 马威；

【导师】 杨胜刚；

【作者基本信息】 湖南大学 ， 金融学， 2013， 博士

【摘要】 金融风险是现代金融学的一个核心问题，它既关系到国家的金融安全，也关系到每个个体和企业的切身利益。特别是近些年来爆发的美国次贷危机和欧洲债务危机，更加引起了学者们对金融风险的广泛关注。而通过构建金融危机预警指数，并对预警指数在中国金融危机时间演化特征、空间关联特征、影响因素和预测等方面的应用，具有重要的意义。首先，文章进行了金融危机预警的基本理论研究。在金融危机相关文献的系统梳理的基础上，将现有文献分为模型预警和指标体系预警两个大类，进而提出金融危机预警指数。在评价了现有文献的基础之上，提出了本文的研究的主要内容与方法、思路和框架。界定了金融风险和金融危机这两个重要的概念，并阐述了金融风险和金融危机之间的关联，提出了警级指数这个概念并作为研究金融危机预警的切入点。从金融危机发生的时间、导火索和原因等方面系统的比较了近二十年来的典型金融危机，并阐述了现有金融危机形成机制，由此概括出中国金融危机的形成机制。在此基础之上，构建了基于历史演进的金融危机预警框架体系。其次，本文提出了基于金融危机预警警级指数的金融危机预警理论。本文对已有金融危机预警模型从基本思想、统计方法、模型适应性等方面进行了全方位的比较，并对这些模型进行了评价。从金融机构、政府部门、企业部门和对外指标四个方面构建了金融危机预警模型的备选指标，利用结构方程模型确认了对金融风险显著性的指标。在对数据进行标准化处理和测算了金融危机预警指标的时滞性之后，通过变异系数、熵值、相关系数和CRITIC四种赋权方法加权合成，得到了金融危机预警警级指数，并简要分析了金融危机预警警级指数的统计性特征。再次，应用金融危机预警警级指数，从时间维度、空间维度和影响因素三个角度分析了我国自2002年7月至2012年12月的金融风险。在时间维度上，从金融危机预警警级指数的基本统计特征、趋势性特征和阶段性特征三个方面，逐层深化地分析了我国金融危机预警警级指数的时间演化特征。研究表明，从2002年7月到2012年12月之间，中国的金融风险呈现明显的趋势性特征，以2007年2月作为拐点，中国的金融危机预警警级指数呈现先增大后减小的趋势。进一步的利用有序样品聚类研究表明，在研究时间区间内，中国的金融危机预警警级指数可以明显的划分为四个不同的阶段，即平稳阶段(2002年7月至2006年2月)、上升阶段(2006年3月至2007年7月)、持续高位阶段(2007年8月至2011年8月)和下降阶段(2011年9月至2012年12月)等四个阶段。在空间维度上，本文从金融风险传染效应和中国金融危机环境出发，阐述研究金融风险预警空间关联性的必要性和重要性。然后构建了灰色关联度模型，从汇率指数和股票指数两个方面，分四个阶段，计算了中国同美国、德国、英国和日本等主要经济体之间的金融风险的灰色关联度，并对结果进行了分析。结果表明：在股票指数方面，中国同四国之间的金融危机的灰色关联性没有显著的地域性差异，但是在四个国家中，均有明显的阶段性差异。在汇率指数方面，中国同欧元对美元汇率的灰色关联度明显低于其他三国的灰色关联度，并且四个国家都表现出了明显的阶段性差异。利用状态空间模型分析了影响金融危机警级指数的因素。对影响因素分析表明，信贷总额、政府赤字额和CPI等三个因素对于金融危机预警警级指数有着显著性的影响，并且在不同阶段表现出影响大小不同，在某些特殊的阶段还呈现出方向性的差异。最后，本文应用已经建立起来的金融危机预警的理论，从点预测和区间预测的角度，分别利用GM(1,1)模型和状态空间模型预测了我国近期的金融危机预警警级指数，并对预测的结果进行了分析。在文章的结尾，利用本文提出的理论，针对中国目前的现状，提出了防范政府债务风险、缓解结构性失衡、防范股市异常波动、降低汇率风险等对策和建议，并对本文研究的内容进行了总结和展望。更多还原

【Abstract】 Financial risk is a core problem of modern finance, which is not only related tothe national financial security, but also related to the vital interests of each individualand enterprise. Especially in recent years, the outbreak of the American subprimemortgage crisis and the European debt crisis, have aroused extensive concern ofscholars on the financial risk. Meanwhile, it’s of essential significance to establishfinancial crisis warning index,and apply it to study evolution characteristics, spatialcorrelation characteristics, influencing factors and prediction of China’s financialcrisis.First of all, this paper studies the basic theory of the early warning of financialcrisis. On the basis of sorting relevant study on the financial crisis, it divides theexisting literature into two categories: the study on the early warning model and thestudy on the index system of the early warning, and then puts forwards the concept ofthe financial crisis warning index. After evaluation of the existing literature, the maincontent and methods, train of thoughts and framework of the research are proposed.First, it defines the two important concepts of financial risk and financial crisis, andexplains the relationship between them. On this basis, the early warning of financialrisk has been established as the core content of this paper, and it puts forward theconcept of alarm index as a starting point for the study of early warning of financialrisk. Then, it systematically compares typical theories on financial crisis in the pasttwenty years in terms of occurrence time, the fuse and reasons, describes formationmechanism of current financial crisis, and generalizes the formation mechanism of thefinancial crisis in china. Finally, it builds the framework of the early warning offinancial crisis, based on historical evolution.Secondly, this paper puts forward the theory of early warning of the financialcrisis which is based on financial crisis warning alarm index. First it makes acomprehensive comparison and evaluation in existing financial crisis early warningmodels from the view of basic theory, statistical method and model adaptation, selectsalternative indicators from four aspects of financial institutions, government sector,non-financial corporate sector and external indicators, and uses structural equationmodel to confirm indicators sensitive to financial risk. After standardizing thecollected data and calculating the delay of early warning indicators of financial crisis,the financial crisis warning index is acquired through weighted synthesis of four weighing methods involving the variation coefficient, entropy, correlation coefficientand CRITI, besides, its statistical characteristics has been briefly analyzed.Furthermore, the financial crisis warning index is applied in analysizing thefinancial risk of China from July2002to December2012from three angles of timedimension, space dimension and the influence factors. In time dimension, this papersystematically describes the time evolution characteristics of the financial cisiswarning index from three aspects of its basic statistical characteristics, trendcharacteristic and periodical characteristics. Study shows, from July in2002toDecember in2012, China’s financial risk presented obvious trend characteristics. TheFebruary in2007was the turning point, in which, China’s financial risk early warningalarm index increased at first and then decreased. Further study using clusteringordered samples shows that, in the study period, China’s financial risk early warningalarm index can be obviously divided into four stages, namely, stable stage(July2002to February2006), ascendant stage(March2006to July2007), sustained high stage(August2007to August2011)and descendant stage(September2011to December2012).In space dimension, in the view of the financial risk contagion effect and thefinancial crisis environment of China, this paper clarifies the necessity andimportance of study on financial risk warning spatial correlation, and constructs thegrey correlation model. The grey correlation of the financial risk between China andthe major economies like the United States, Germany, Britain and Japan is calculatedfrom two aspects of exchange rate and stock index in four stages, and the results areanalyzed. The results show that, in the asepect of stock index, there is no significantregional discrepanies in the gray correlation of the financial crisis between the abovefour countries and China, but there are obvious periodical difference in each of thefour countries. As for the exchange rate, the gray corrcletion between China’swarning index and the exchage rate of euro against the dollar is weaker than that inother three countries,in addition, all those four countries show obvious periodicaldifference. A state space model is used to analysis the factors influencing the financialcrisis warning index. Factor analysis on these factors shows that, three factors,thetotal credit, government deficits and the CPI, have a significant influence on financialrisk early warning alarm index, and influence is different at different stages, in somespecial stages showing directional differences.Finally, the established financial crisis early warning theory is applied, alongwith the use of GM (1,1) model and the state space model, respectively, in the view of point prediction and interval prediction, to predict China’s recent financial risk earlywarning alarm index, and the prediction results are analyzed.At the end of the article,based on the theory presented in this paper, according to the current situation in China,measures and suggestions on financial risk prevention and solution are put forward.Meanwhile, a summary and expectation on the study subject are presented.更多还原