论文部分内容阅读
运用国内商业银行积累的大量数据,统计得到银行个人客户住房抵押贷款多年度、不同信用等级、不同身份特征、分行业和分地区的违约情况,进行非线性的拟合分析,并采用Copula函数度量个人客户违约之间的相关性及厚尾特征。研究表明,房屋价格、客户性别以及受教育程度等与违约概率相关性比较低,在考察的样本区间内,这些因素不显著导致违约发生。另外,信用等级、收入结构和抵押担保剩余额度是影响个人违约决策的重要变量。所采用的模型在个人住房抵押贷款定价与风险管理中获得较好效果,银行可以根据违约状况的变动制定动态利率,随时准备弥补损失。
By using the large amount of data accumulated by domestic commercial banks, we can obtain the non-linear fitting analysis of the default of mortgage-bearing bank customers for many years, different credit ratings, different identities, industries and sub-regions, and use the Copula function to measure Correlation between individual customer default and thick tail features. The research shows that the correlation between house price, client’s gender and education level and the probability of default is relatively low. These factors do not significantly lead to default during the sample interval. In addition, the credit rating, income structure and mortgage guarantee residual amount are important variables that affect the individual’s default decision. The model used in the individual housing mortgage loans pricing and risk management to obtain better results, the bank can make changes in the status of default dynamic interest rates, ready to make up for losses.