【作者】 朱小宗；

【导师】 张宗益；

【作者基本信息】 重庆大学 ， 技术经济及管理， 2005， 博士

【摘要】 在现代市场经济中,金融体系是经济体系的核心,而银行是金融体系的重要组成部分,因此,银行的运行质量和效率已成为经济发达程度的一个重要标志。但是,银行依存于经济环境,要受到国家经济体制、国家经济政策、法律制度、经济发展水平、企业经营状况等方面的影响,面临着信用风险、市场风险和操作风险,但最主要的风险还是信用风险。正是信用风险在银行经营和管理中极其重要性,巴塞尔委员会从1988 年到2004 年先后出台了一系列关于加强银行包括信用风险内的风险计量、管理和控制的指导文件,国际性大型商业银行和研究机构也对信用风险度量和管理提出了新的理论和计量方法,并进行了大量的实践活动。由于我国银行业的改革滞后于经济发展,在风险度量和管理方面还很薄弱,因而,研究和发展风险度量模型和方法对我国银行业具有十分重要的意义。本文研究得到的创新点概括如下: 一、全面剖析了传统的和现代的信用风险度量模型,并进行了实证比较研究。利用我国银行的大量信贷数据对信用风险度量的传统模型和现代模型进行实证研究,进行综合比较分析,分析了在我国的适用性;并基于银行信贷数据得出我国企业信用等级迁移矩阵、死亡率表、各信用等级下的违约率和违约下的损失率及其分布等,这些相比国内主要是介绍国外理论和现代模型要前进了一步。研究认为传统模型在我国基本上不具有适用性;现代模型测算结果差异性很大,而且不能很好地测算单一信贷资产损失,大多数模型测算的组合贷款损失率大于我国银行贷款资本配置比例的实际值(8.24%),但巴塞尔委员会的要求值(8%)小于实际值(8.24%);我国银行企业实际违约率大于巴塞尔委员会的推荐值。研究还得出,我国银行对企业的信用评级相比穆迪公司和标准普尔的信用评级具有连续性差、波动性大、迁移不平稳、违约率高的缺陷;对银行贷款的PD、LGD 进行分类统计应主要按信用等级来统计,在此基础上按行业性统计可以提高统计值的准确度;西方国家开发的模型往往直接或者间接地认为PD 和LGD 相互独立,但笔者样本统计表明两者存在强烈的正相关性。二、建立了度量信用风险的新模型,并与现有的模型进行范式和实证比较。笔者的规范研究和实证研究表明,借款人偿债能力越低,越容易产生道德风险,也容易产生信用风险,两者产生的机理和临界条件是完全相同的,影响这种行为最根本的原因是企业的产出效率,只要企业产出效率超过临界值,那么企业更多还原

【Abstract】 In modern market economy system, the finance system is a core to the economy system, and banking is an important composition of the system. So, the performance quality and operation efficiency of bank institution is an important sign to the extent of economy development level. However, its development depends on macro-environment, and it will be influenced by migration of the nation economy system, economy policies, law institution, level of economy development, operation performance of enterprises, and so on. Banks are faced with credit risk, market risk and operation risk, but the credit risk is the greatest among them. In view of the importance of credit risk for operation and management in banks, Basel Committee has successively promulgated series of documents on calculating, managing and controlling banks risk included credit risk since year 1988. International large business banks and research institutions had put forward new theories and methods of calculating and managing credit risk and have put them in practice. Because of banking reform lagging economy development in our country, the ability of risk measurement and management in banking is vulnerable. So it is great important for China banking to refer and develop modern risk measurement models and methods. The main contributions are drawn up as follows. Firstly, giving an anatomy and comparative analysis on traditional and modern credit risk measurement models, besides empirical study. By means of much credit data from China banks, the author gives an empirical study on traditional and modern credit risk measurement models, besides an analysis of their applicability for China. Meanwhile, based on the credit data, the paper finds Chinese enterprises’rating migration matrix, mortality ratio table, default possibilities and loss at given default under sorts of rating, which goes ahead of what introduces currently models in China. The study finds that the traditional models have no applicability for China. The forecast outcome of current models is very distinct, and they can’t forecast loss of single loan. The forecast of loan loss proportions of the majority models are more than actual proportion of economic capital for China banks’loans (8.24%), which is more than the proportion required by Bale Committee (8%). Moreover, the actual default possibilities of China enterprises are more than what Bale Committee recommends. The study also finds that the enterprises’ratings in China have such drawbacks as weak continuity, big volatility, unstable migration, and high default possibility compared with Moody’s and Standard & Poor’s ratings. Estimating PD and LGD of bank loans by statistical methods should be firstly sorted by enterprises rating, secondly by industry. Moreover, the study finds that PD and LGD have positive correlation not independence. Secondly, newly-establishing a credit measurement model, with mode and empirical compare with existing models. The mode and empirical study indicates that the weaker solvency enterprises, the easier for moral hazard and credit risk occurring, and they have the same mechanism and critical condition, which is output efficiency. When the output efficiency is over critical condition, the default possibility of enterprises decreases greatly. According to the research, the paper newly-establishes a current risk measurement model based on output efficiency. The new model studys on enterprises’default behavior and loans loss based on their financial structure, and the model not only has representative features, but also can give four parameters, which required by Basel Committee. Mode analysis and empirical comparison analysis demonstrate that the new model is superior to credit risk models that the West developed, and that the new model has good applicability in China. Especially, the new model can calculate not only loss of single loan but also loss of portfolio, and it doesn’t require much credit data. Thirdly, putting forward new practical use methods for banks to manage credit risk by credit risk measurement models. The author puts forward new-pricing of loan based on the facts that the pricing method of traditional method (mainly Cost Plus Profit) doesn’t calculate loan’s unexpected loss and that the method of RAROC underprices the loan, so the author puts forward new method of loan pricing at given profit, which involves loans’unexpected loss and return of economic capital-adjusted. The new pricing method can give a section of loan’s prices, which supports for making decision of loans. The new model can give enterprises’simple rating with dumbbell structure. It is necessary to use other information to better rating in practical use. The rating method can give credit line on enterprises, and empirical study displays that the credit line calculated by the model is less than what banks do, because the model involves the loan’s loss and the rating estimated by the model is inferior to actual rating, but it is more practical for banks to control risk.更多还原