Empirical Research On The Optimization Of The Real Estate Development Loan Portfolio Of Commercial Banks In China

In the financial sector, Credit risk has been the most important and themost difficult to measure for commercial banks. The real estate industry withpilot and fundmental characteristics highly associates with the commercialbanks as well as other industries, which is a signal of the national economicdevelopment and also the potential risks of the credit system.The2007U.S."subprime" crisis’s root causes is the the credit risk of thereal estate industry, which leads to the2008global financial crisis. Therefore, itis very important to measure the credit risk of the bank loans of the real estatebusiness. This requires first to measure single housing company’s credit risk,then to acquire the the very important default dependency structure betweendifferent housing companies, and finally to optimize the combination of realestate development loans.From the point of view of the commercial banks, with three A-share listedhousing companies for the study, and based on the stock prices and financialdata from the beginning of2006to the end of the third quarter of2012, the firststep is to calculat the default distance of a single housing company by using theKMV model, the second step is to fit the dependence structure of the defaultdistance of the three samples housing enterprises by using the Copulaconnection function. The research results show that the default distance betweenany two of the three listed housing enterprises has almost the same binaryGumbel-Copula connection function, which means that there exists a certaindependence structure on the upper tail beteen default distance. With the nature of the Gumbel-Copula function, we further construct a ternary Gumbel-Copulafunction as the empirical distribution function to study the dependence structureof the three default distance. Finally, we tentatively use the distance randomnumbers generated by the the empirical distribution function, and define a newmeasure of portfolio credit risk variable “dd” by using the weighted average ofeach set of random numbers. By the way, the weight here is the allocation ratioof the loan. At last, similar to the VaR thinking, through optimizing the loanportfolio we obtain some relatively optimal loan allocation ratio, For example,when the level of significance equal to0.05, Zhongliang,Yukaifa and Dalongwere given about20%,30%,50%of the loan. In addition, when the level ofsignificance rises from0.01to0.1, the proportion of Zhongliang and Yukaifagradually increase, while the proportion of Dalong gradually declines.