【作者】 陈开方；

【导师】 胡则成；

【作者基本信息】 武汉理工大学 ， 企业管理， 2003， 硕士

【摘要】 信用风险是金融市场中最古老也是最重要的金融风险形式之一。它是现代社会经济实体(尤其是金融机构)、投资者和消费者面临的重大问题。分析信用风险形成的原因，从而防范信用风险成为现代金融领域需要研究的一个重要课题。 导致银行信用风险的主要原因是经济主体的信用缺失。信用缺失的基础是有限理性和机会主义的存在。通过对信用缺失的经济学分析，笔者认为防范和控制信用风险首先要从根治失信行为开始。而根治失信行为必须从产权着手，必须抓制度建设；必须引入政府规制并辅之以伦理、道德等非正式的约束机制；必须重视信息的建设、管理和应用。 通过与国际先进的商业银行信用风险管理体系对比，并借鉴国际先进商业银行风险管理的经验，提出了理想的风险管理体系必须牢固建立在制度和技术两个平台之上，并将商业银行信用文化建设和外部信用环境建设纳入其中。风险管理的战略、偏好、政策、过程及组织架构均可以认为是风险管理体系制度要素。针对目前国有商业银行风险管理的“软肋”恰恰就在于技术平台的薄弱，介绍了目前国际银行界流行的资产组合管理模型，并进行了分类。指出实现信用风险管理的终极目标——总体权衡风险与收益之间关系的资产组合管理——依然应以“客户评级体系”和“贷款风险分类”构成的二维评级体系为风险识别的重要内容，并从单笔资产的风险识别开始。 出于评价客户信用等级的目的，介绍了信用风险识别最著名的Z模型。针对Z模型的不足，指出应结合实际对客户评级体系进行修正，并提出了客观评级的校正标尺。 根据巴塞尔新资本协议，结合我国国有银行业目前贷款风险分类的现状，指出五级分类需要进一步细化，以便清楚地反映贷款质量的变化过程，真正实现根据贷款质量定价的“优质优价”策略。 除了规范性的定性分析之外，本文还对银行信用风险的预警机制和预测模型进行尝试性的研究。参考Bosman A. Roger提出的预测贷款违约主要先期迹象和指标，综合影响商业银行信用风险预警基本因素，遴选出一些关键指标建立银行单笔贷款信用风险的预测模型，这个模型在商业银行控制信用风险的实际操作中具有一定的参考价值。更多还原

【Abstract】 Credit risk is one of the most ancient and most important financial risk forms in the financial market, and a major issue facing modern economic body (especially financial institutions), investors and consumers. It has been an issue of great importance in the financial filed of modern economy to analyze its causes in order to guard against credit risk.The main reason leading to credit risk of banks is the shortage and loss of credit in economic bodies. The basis of the shortage and loss of credit is the existence of limited rationalism and opportunism. By analyzing the shortage and loss of credit by means of economics, it is believed in this essay that we should start to guard against and control credit risk through uprooting the behaviour of credibility loss. To uproot the behavior of credibility loss, we can start from property right. It is necessary that we should attach importance to the setting up of rules and regulations, the introduction of biding system of ethics and moral into government laws and rules. We should also attach importance to the construction, administration and application of information.By comparison with the world advanced administrative system of commercial bank credit risk and drawing on the experience of administration of international advanced commercial banks, the essay puts forward the concept that the ideal administration of credit risk should be firmly based on regulations and technology, and construction of culture and external credit environment should also be included. Considering that the "soft rib" of risk administration of state-owned commercial banks lies in the poor condition of technology, the essay introduces and streams the administrative molds of capital and property formation of current world banks. To realize the ultimate goal of the administration of credit risk, namely, to balance the relationship between risk and profit of the administration of capital and property formation, the grading system such as "grading system of clients" and "classification of loan risks" become the important contents of risk identification, which starts from risk identification of single property risk.For the purpose of evaluating the grades of clients’ credibility, the essay introduces the most famous Z model. Concerning the shortage of Z model, the essay points out the necessity of modification of clients grading system in a practical way, and offers the standards of modification of practical grading.According to the Bassel’s agreement of new capital, and combined with the statue quo of the classification of loan risk of our state-owned banks, it is pointed out that the five levels of classification should be further streamed so that the changing process of loan quality can be clearly reflected and the tactics of Good Quality and Good Price which is priced according to the loan quality can be achieved.Except for the standardized qualitative analysis, the essay also makes attempted research into the forewarning system and prediction models of bank credit risk. In reference with the major earlier symptoms and index of agreement violation of loan put forward by Bosnian A. Roger, and combining the basic elements that influence credit risk forewarning of commercial banks, some key indexes have been selected to establish the prediction model of single loan credit risk. The model may serve as reference in the practice of controlling credit risk of commercial banks.更多还原