Research On Prepayment Prediction Of Commercial Bank Mortgage Based On Survival Analysis

Mortgage prepayment will put additional strain on the allocation of credit funds of commercial banks,resulting in a loss of expected interest income.Its adverse effects are increasingly becoming more noticeable,which are primarily concerned with the expansion of the mortgage scale.The People’s Bank of China implemented loan LPR interest rate pricing in 2019,which also caused undetermined changes in the interest income losses of commercial banks’ mortgage loans.In order to explore the variable factors impacting the occurrence of mortgage prepayment and to analyse the impact of various variable factors on the duration of the mortgage,particularly for quantifying the impact of the LPR interest rate pricing model,for predicting the prepayment of mortgage loans of commercial banks,for solving the liquidity risk caused by the cash flow of prepayment of mortgage loans,and for solving the loss of expected interest income caused by the prepayment of mortgage loans under the LPR pricing model.The survival analysis approach and its application model,the Cox Proportional Hazard Model,are used in this paper to investigate the prepayment prediction of commercial banks.After giving a comprehensive review of the method and model of mortgage prepayment prediction,the survival analysis theory was chosen and the Cox proportional hazards regression model,which is a semiparametric model of survival analysis was used in this study,from data on 6895 mortgage loans of a city commercial Bank A so as to understan d the prepayment performance,given a variety of factors that influence the prepayment.The survival analysis method is employed to explore the connection with birth-death processes,study the duration of occurrence of specific events and the factors impacting the occurrence of specific events,therefore this method is applicable to describe the duration of the mortgage as well as its influencing factors.The Cox proportional risk model can be utilized to determine the probability of a specific event,particularly for the large-sample problem with an unknown duration distribution.Cox regression model with covariates can be used to describe the duration of the mortgage lifetime and its prepayment,whereas the mortgage life function is unknown.The statistica l significance of the indicators was verified using the Kaplan-Meier single-factor test,and the log-rank test was used to compare different levels.For regression analysis of the five categories of indicators that passed the test,a Cox model was construc ted.Wald test was utilised to estimate the partial regression coefficient.The loan survival curve is made,the prepayment prediction model is obtained,and the probability of settlement outcome is described by the risk ratio HR.According to mortgage survival data,the descriptive statistical analysis of the life time of the mortgage reveals that the prepayment behaviour is the highest when the mortgage term is about half.In the case of a large sample,the distribution of mortgage life time asymptotic follows the normal distribution.After conducting Cox model regression dynamic analysis,which indicates LPR interest plus a key variable,the more loan interest rate LPR point increase,the more likely mortgage to prepayment,and the prediction model demonstrates that LPR point has a major impact on the outcome of prepayment,LPR point increased by 1 unit,and the loan settlement risk increased by 3 times.Finally,based on the findings of the prepayment prediction model,the paper puts forward suggestions for commercial banks to adopt different prepayment responses for customers with varied prepayment risks in prepayment management.In addition,the conclusion of the mortgage prepayment forecast can provide data support for the pricing of Residential Mortgage Backed Security products.