The Influencing Factors And Prediction Research On Housing Mortgage Prepayment

After the financial crisis in 2008, the benchmark lending rates experienced a decline. However, from 2010 to now, the central bank frequently used monetary policy, raised interest rates four times. According to the exercise between 2004 to 2007, as benchmark lending rates rising, the mortgage prepayment appeared in large Numbers. The recent tightening monetary policy in preventing inflation also may lead the banks facing peak prepayments again. Prepayment make the banks encounter expected interest income losses. Therefore, banks needs to explore factors which influence the prepayment of borrowers urgently, in order to maximize returns.The purpose of this paper is to explore the factors influence mortgage prepayment, and give suggestions to banks, based on the conclusion. At the same time, establish predicting model to predict borrowers'prepay actions. According to the predicted results, banks can adjust their assets and liabilities structure to meet borrowers'prepayment, avoiding sudden prepayment lead to expected interest income loss.In order to achieve the research purpose, this thesis based on the SPSS 18.0, used 53218 valid samples collected to do the factor analysis at first. And extracted factors according to the principle of characteristic value greater than 1, then extract factors which have significant influences to prepayment according to the principle of the factor loading coefficient exceeds 0.5, Then used the extracted factors to do gradually logistic regression analysis. Secondly, based on the 53218 valid samples collected to do the discriminate analysis, established discriminate function to predict the future state of the occurred mortgage loan, and established the Fisher discriminate function for simple discriminate; Still did discriminate analysis in different housing types, housing position and loan types,in order to predict the future payments state more concretely. This thesis base on quantitative analysis mainly and qualitative analysis secondarily. Qualitative analysis was used to analyze the reason of choosing to prepay and the impact of prepayment on banks, quantitative analysis was used to analyze the factors that affect prepayment and to predict the future prepayment.Based on the above ideas, this thesis is divided into five chapters, specific arrangement is as follows:In the first chapter, the background, significance, research ideas and the structure arrangement of this thesis are introduced. The research variables and research methods about mortgage prepayment are summarized, after summarizing the existing research at home and abroad.Chapter two focused on the general analysis of mortgage prepayment. The reason of borrowers'choosing to prepay, the impact of prepayment to banks and the risk management of prepayment are analyzed in this chapter.Chapter three concerned the Statistics describing analysis of the mortgage prepayment. The variables characteristics of the prepaid sample was understood by comparing statistical features of the normal samples, prepaid samples and the overall samples. Meanwhile, Analyzed the diachronic characteristics and the interest spread characteristics of the prepaid samples, found that the average time of duration is 26.16 months, and as long as interest spreads to a certain extent numerous prepayment appeared.Chapter four is the most important section of this thesis; this chapter is divided into three parts.The first part emphasized on the empirical analysis of the factors affecting prepayment. First, extracted seven factors, according to the principle of characteristic value greater than one, through the factor analysis. Second, deleted the borrower gender, household registration and degree, according to the principle of load coefficient less than 0.5. Finally, did the gradually logistic regression analysis with the remain seventeen factors, age and total house price variables are deleted from the logistic regression model. The empirical results show that the borrowers who are single and monthly income is higher select prepay easily; The mortgages with houses which located in the downtown, with higher unit price and wide area are easy to be prepaid; The mortgages with short term, higher rate, less amount and more initial payment are easy to be prepaid; The mortgages with higher prices index lower CPI are easier to be prepaid.The second part is the discriminate forecast analysis. This chapter in combination with factor analysis, deleted the borrowers'gender, household registration and degree, established the gradually discriminate analysis about 16 variables (marital status, monthly income, unit prices, total prices, housing types, housing position, housing area, loan types, loan periods, initial loan interest rates, mortgage amount, monthly debt revenue ratio, and housing price-income ratio, house price index and CPI).The discriminate function has a prediction accuracy of 70 percent to the normal samples and 70.5 percent to the prepayment samples.The third part is the classified discriminate analysis. According to the conclusion that housing types, housing position and loan types have significant influence to prepayment in part one and part two, this part established classification discriminate function and Fisher typical discriminate function according to different housing types, housing position and loan types. The classified discriminate function can forecast the future state of the mortgage more concretely.Chapter five is the suggestions part of this thesis. This chapter puts forward suggestions from the point of bank credit review and risk management.This paper selected 83935 mortgage samples occurred between 2004 January 1 to December 31,2009, and got 53218 valid samples. The sample size is of the largest in the existing domestic research. In reference to the previous studies, this thesis used factor analysis, logistic regression analysis and discriminate analysis. Still established the classified discriminate function, which is rare in the existing research, belongs to the innovation of the contents. The conclusion that when interest rates spreads to a certain extent, positive or negative interest rates would lead to substantial prepayment, is the innovation of this thesis on the conclusive aspect.