基于Logistic模型的信托贷款信用风险管理研究

发布时间：2018-11-14 07:32

【摘要】：在中国,信托公司是唯一能够同时涉足货币市场、资本市场和产业市场的“三栖”金融机构。从基础设施、大型工程建设投融资到租赁、担保到企业的兼并重组、改制顾问,信托公司能够提供全程式的金融服务。美国经济学家斯考特这样评价信托：信托的应用范围,可以和人类的想象力相媲美。信托机构作为管理财富的企业,在为社会提供各种金融服务的同时,也面临着市场风险、操作风险、信用风险等各种形式的风险,其中信用风险是最主要、最直接的风险,与其他种类的风险相比也更具有不可控制性。本文总体思路是认真梳理国内外信用风险管理的研究现状和研究成果,为信托贷款信用风险的分析搭建理论平台,剖析信托贷款的信用风险,探求信托贷款信用风险管理的有效方法。首先,总结比较几种信用风险管理的传统方法和现代模型的优缺点,结合我国实际情况,最终选择logistic回归模型作为本文的实证分析工具；然后,采用139家申请信托贷款企业的相关数据与资料作为样本,寻找影响信托贷款信用风险的真正因素,建立信托贷款信用风险logistic回归模型,再用该模型检验A信托贷款项目的违约可能性。结果表明,信托贷款信用风险Logistic回归模型是有效可行的；最后,结合我国当前的经济政策,提出信托公司防范和治理信托贷款信用风险的对策和建议。第一部分绪论。介绍了研究背景、研究目的及意义、国内外研究现状、研究方法与思路、研究创新与特色。第二部分信托贷款信用风险管理概述。首先,介绍了信用风险的含义和特点；其次,从管理的角度阐述了信托公司信用风险；最后,从管理的角度阐述了信托贷款信用风险。第三部分信托贷款信用风险度量方法。首先,介绍了几种传统的信用风险管理方法；其次,介绍了几个比较有影响的信用风险管理模型；通过对比各自优缺点,最后指出Logistic回归模型比较适合用来度量信托贷款信用风险。第四部分是Logistic模型在信托贷款风险评估中的实证分析。首先,介绍了Logistic函数如何转化为模型；其次,阐述了模型指标和样本数据的选取,指出财务指标和非财务指标的选取标准和意义,样本数据的来源；最后,构建了信托贷款信用风险Logistic模型。第五部分是基于A信托贷款项目的Logistic模型实证检验。以A信托贷款项目为例,验证信托贷款信用风险Logistic回归模型的有效性,表明信托贷款信用风险Logistic回归模型有较强的可行性。第六部分结论和建议指出,西方国家现代信用风险管理技术已经比较成熟和完善,但目前我们利用西方信用风险管理模式的条件还不具备,我国金融行业还需要在数据采集、发展商业信用评级机构和完善相关法律、体制方面加强防范措施。
[Abstract]:In China, trust companies are the only three-perch financial institutions that can tap into money, capital and industrial markets at the same time. From infrastructure, investment and financing for large projects to leasing, guarantees to mergers and reorganization of enterprises, restructuring consultants, trust companies can provide full-course financial services. The American economist Scott said of trust: trust can be applied in a range comparable to human imagination. As an enterprise managing wealth, the trust institution is facing various kinds of risks, such as market risk, operation risk, credit risk and so on, while providing various financial services to the society. Among them, the credit risk is the most important and direct risk. They are also more uncontrollable than other types of risk. The overall idea of this paper is to carefully comb the domestic and foreign credit risk management research status and research results, for trust loan credit risk analysis of the theoretical platform, analysis of trust loan credit risk, To explore the effective method of trust loan credit risk management. Firstly, the advantages and disadvantages of several traditional methods and modern models of credit risk management are summarized and compared. Finally, logistic regression model is chosen as the empirical analysis tool of this paper. Then, the logistic regression model of trust loan credit risk is established by using the relevant data and data of 139 trust loan enterprises as samples to find the real factors that affect trust loan credit risk. Then the model is used to test the possibility of default of A trust loan project. The results show that the Logistic regression model of trust loan credit risk is effective and feasible, and finally, combined with the current economic policy of our country, the countermeasures and suggestions for trust company to prevent and control trust loan credit risk are put forward. The first part is introduction. This paper introduces the research background, research purpose and significance, domestic and international research status, research methods and ideas, research innovation and characteristics. The second part is an overview of trust loan credit risk management. Firstly, it introduces the meaning and characteristics of credit risk; secondly, it expounds trust company credit risk from the perspective of management; finally, it expounds trust loan credit risk from the angle of management. The third part is the credit risk measurement method of trust loan. Firstly, several traditional credit risk management methods are introduced, secondly, several influential credit risk management models are introduced. By comparing their advantages and disadvantages, it is pointed out that Logistic regression model is more suitable to measure trust loan credit risk. The fourth part is the empirical analysis of Logistic model in trust loan risk assessment. Firstly, it introduces how to transform Logistic function into model, secondly, expounds the selection of model index and sample data, points out the standard and significance of selecting financial index and non-financial index, and the source of sample data. Finally, the Logistic model of trust loan credit risk is constructed. The fifth part is the empirical test of Logistic model based on A trust loan project. Taking A trust loan project as an example, the validity of the trust loan credit risk Logistic regression model is verified, which indicates that the trust loan credit risk Logistic regression model has strong feasibility. The sixth part points out that the modern credit risk management technology in western countries has been mature and perfect, but at present we do not have the conditions to use the western credit risk management model, and the financial industry in our country still needs to collect data. To develop commercial credit rating agencies and improve relevant laws, institutional measures to strengthen prevention.
【学位授予单位】：安徽大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F832.49

【参考文献】