基于Copula函数的商业银行整合风险度量

发布时间：2018-03-25 00:37

本文选题：整合风险　切入点：商业银行　出处：《河南师范大学》2014年硕士论文

【摘要】：20世纪80年代以前，商业银行主要是以单一、局部的管理方式来管理其所面对的金融风险。90年代以来，随着金融技术的进步，经济全球化趋势的迅速蔓延，两个非常严峻的现实摆在了银行风险管理的面前。它们分别是：一，，经济的全球化速度加快以及行业集团化的现象更加明显；二，信息计算科学的广泛使用。这些现实都使得商业银行所面对的风险由单一的信用风险走向多样化、复杂化。随着巴塞尔新资本协议的颁布，商业银行的风险管理模式也随之改变，即从以往的单一的信用风险管理模式转变为信用风险、市场风险、操作风险三种风险并举的全面的风险管理模式，同时指出各风险之间存在着某些相关性[1]。既然多种风险之间存在着相关性，那么银行在风险管理方面不得不制定出更完善的应对策略。混业化的经营是这世纪所有金融行业的发展趋势。在这种大的经济背景下，银行业更是在混业经营方面表现最为突出的一个行业。由于不同的业务资本的组合而产生的混合风险使得单一的风险管理模式无法应对，因此将各类风险进行整合起来管理越来越引起研究者和风险监管者的重视。因此，整合风险管理也成为当代商业银行风险管理的最为突出的发展趋势。本文首先对商业银行所面对的三种风险进行了概括性的分析，分别介绍了它们各自的金融学含义、特征并且简明扼要的总结了实践中常用的度量方法。然后简单介绍了VaR、CVaR，Copula函数的相关内容以及Monte Carlo模拟的基本步骤。在过去的很长一段时间里，大多数的研究者和风险监管者使用的都是将不同风险值简单的相加，以用来估计总体风险，但是这种方法却大大的高估了整合的风险值。基于Copula函数的整合风险的度量模型可以更加准确的描述不同风险边际分布间的相依结构，那么对风险的度量模型不再是对单个的风险的度量，而是整体考虑了金融机构所面对的三种主要风险。本文在整合度量商业银行所面对的三种风险时采用了Copula函数和CVaR方法。同时分别用Normal copula和t-copula函数来描述三种风险之间的相依结构，同时构造了它们的联合分布函数，最后通过Monte Carlo模拟来估计不同的相依结构下整合风险的CVaR值。通过此模型可以计算出商业银行所面对的这三种主要风险的的整合CVaR值和VaR值，从而可以比较出两者的优劣，同时本文也研究了不同的权重组合CVaR值的影响。
[Abstract]:Before the 1980s, commercial banks managed the financial risks they faced mainly in a single and partial way. Since the 1990s, with the development of financial technology, the trend of economic globalization has spread rapidly. Two very serious realities are facing the risk management of banks. They are: first, the rapid pace of economic globalization and the phenomenon of industry collectivization is more obvious; second, With the widespread use of information computing science, these realities have made the risks faced by commercial banks diversified and complicated from a single credit risk. With the promulgation of the New Basel Capital Accord, The risk management mode of commercial banks has also changed, that is, from the former single credit risk management mode to the comprehensive risk management mode of credit risk, market risk and operational risk. At the same time, it is pointed out that there is some correlation between various risks [1]. Well, banks have to work out a more complete coping strategy in risk management. Mixed operation is the development trend of all financial industries in this century. Under such a large economic background, The banking industry is the most outstanding industry in the aspect of mixed operation. Because of the mixed risk caused by different portfolio of business capital, a single risk management model can not cope with it. Therefore, the integration of various types of risk management has attracted more and more attention of researchers and risk regulators. Therefore, integrated risk management has become the most prominent development trend of risk management in contemporary commercial banks. In this paper, three kinds of risks faced by commercial banks are analyzed, and their respective financial meanings are introduced respectively. Features and a brief summary of the commonly used measurement methods in practice. Then, this paper briefly introduces the relevant contents of the Monte Carlo function and the basic steps of Monte Carlo simulation. In the past a long time, Most researchers and risk regulators use simple sums of different risk values to estimate the overall risk. However, this method greatly overestimates the risk value of integration. The model based on Copula function can more accurately describe the dependent structure of the marginal distribution of different risks. So the risk measurement model is no longer a measure of a single risk, In this paper, Copula function and CVaR method are used to measure the three kinds of risks faced by commercial banks. Normal copula and t-copula function are used to describe the three kinds of wind, respectively. The dependent structure between risks, At the same time, the joint distribution function is constructed, and the CVaR value of integration risk under different dependent structures is estimated by Monte Carlo simulation. By this model, the integrated CVaR value and VaR value of the three main risks faced by commercial banks can be calculated. So we can compare the advantages and disadvantages of the two. At the same time, this paper also studies the influence of different weight combination CVaR value.
【学位授予单位】：河南师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F830.3

【参考文献】