基于生存分析法的住房抵押贷款违约风险影响因素分析

发布时间：2017-12-27 17:37

本文关键词：基于生存分析法的住房抵押贷款违约风险影响因素分析　出处：《湖南大学》2016年硕士论文　论文类型：学位论文

【摘要】：美国次贷危机发端于银行次级住房按揭贷款爆发高违约率,如果银行可以对该类贷款的违约概率做出准确预测,则能使危机爆发的可能性大大降低。我国的商业银行现今存在大量的住房抵押贷款,如何有效识别和度量其违约风险从而防止此类危机的爆发,对我国商业银行而言是一个重要挑战。本文旨在为我国商业银行住房抵押贷款违约概率的测算提供一种有效的方法。传统的简约模型和现代结构模型都能一定程度测算住房抵押贷款违约概率,然而由于住房抵押贷款期限较长,观察期内结束的贷款样本较小,而以上两种方法都直接删除了未完成贷款的样本,这就会导致样本量过少以及信息缺失。生存分析方法最大的优点在于考虑了删失数据,提高了模型测算的准确度。本文回顾了住房抵押贷款风险及其影响因素和生存分析方法的基本理论,并基于Cox模型,从借款人特征、贷款特征、房屋特征三个维度对住房抵押贷款违约风险的影响因素进行了计量分析,模型拟合度较好,计量结果较为显著,最终得出月还款额占月收入比、学历水平、婚姻状况、利率、首付比例、月还款额、贷款期限都对住房抵押贷款违约风险有显著影响,其中月还款额占月收入比的影响最大。同时,本文对提前还款的样本进行了分析,得出了借款人性别、学历水平、利率、房屋面积对借款人是否会提前还款影响较大的结论。本文运用竞争风险模型分析了违约、提前还款和正常还款三者之间的竞争风险,得出违约与提前还款以及正常还款的影响因素不同,因此在测度住房抵押贷款违约概率时,将提前还款和正常还款作为删失数据是适宜的。而提前还款和正常还款之间则存在竞争风险,对正常还款和提前还款而言,学历水平、月还款额、月还款额占收入比、贷款期限、房屋价值以及首付比例这几个影响因素的系数有非常显著的差别。最后本文根据实证结果给出了针对性的政策建议。
[Abstract]:The US subprime mortgage crisis started from the high default rate of the secondary housing mortgage loan of banks. If banks can accurately predict the default probability of such loans, the possibility of the outbreak of the crisis will be greatly reduced. There are lots of housing mortgage loans in China's commercial banks. How to effectively identify and measure their default risk and prevent such crises is an important challenge for Chinese commercial banks. The purpose of this paper is to provide an effective method for calculating the default probability of housing mortgage loan in China's commercial banks. A simple model of modern and traditional structure model can be a certain degree of probability of default estimates of housing mortgage loans, but due to the longer mortgage period, the end of the observation period of the loan of a small sample, and the above two methods are directly delete the unfinished loan samples, which will lead to small sample size and lack of information. The greatest advantage of the survival analysis method is that the censored data is considered, and the accuracy of the model calculation is improved. This paper reviews the housing mortgage loan risk and its influence factors and basic theory of survival analysis method, and based on the Cox model, the quantitative analysis of impact factors from three dimensions of the borrower characteristics, loan characteristics, characteristics of the housing mortgage default risk, model better fit, the measurement result is more significant, the final monthly payments total monthly income ratio, education level, marital status, interest rates, the proportion of down payment, monthly payments, loans for housing mortgage loan default risk has a significant impact, the monthly repayment amount of monthly income than the maximum. At the same time, this paper analyzes the sample of early repayment, and concludes that borrowers' gender, education level, interest rate and housing area have great influence on whether the borrower will make early repayment. This paper analyzes between default, prepayment and normal repayment three competing risk using competing risks model, factors affecting the default and prepayment and normal repayment, so the measure of housing mortgage loan default probability, the repayment and the normal repayment as censored data is appropriate. Early repayment and repayment is between the normal competition risk, the normal repayment and repayment, there were significant differences in the coefficient of education level, monthly payment, the monthly repayment amount of income ratio, loan period, the value of housing Shoufu ratio and the several influence factors. Finally, this paper gives the pertinent policy suggestions according to the empirical results.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F832.479

【相似文献】