贷款组合风险优化决策模型的研究

Research on the Optimization Decision-Making Models for Loan Portfolio Risk is the project supported by the National Natural Science Foundation of China (NSFC) and the Canadian International Development Agency (CIDA) as the China-Canada University Industry Partnership Program (CCUIPP). Research orientation of the project focuses on the optimal decision-making methods for loan portfolio risk based on risk control. A series of methods and models are set up for loan portfolio on basis of analyzing the characters and defects of similar study both at home and abroad. .(1) Optimal model of asset-liability-management based on the constraint of VaR technology is set up. .Based on the CreditMetrics method and asset-liability-management technology, considering the constrain on VaR, laws, regulations, and operation, using portfolio profits maximum of bank's assets as objective function, the optimal model of asset-liability-management based on VaR technology is set up, in order to provide decision-making method for bank's risk management..The characteristics lies on three aspects: Firstly, by using VaR to set up constrain, and through controlling maximum limitation of loss under certain confidence, the risk of loan distribution is limited within given ranges of bank's risk tolerance ability and reserve funds. Secondly, Asset Liability Management ratios are used to establish constrains and liquidity risk is controlled by using constrains on laws, regulations, and operation, so that the loan's allocation can meet requirement of supervision and operation. Thirdly, the loan's yields of historical data on individual enterprise are used to get the correlation coefficient between different loans, instead of using firm's correlation coefficient to solve the problem as current study, and to get portfolio deviation, thus, the yields' correlation among different loan is reflected directly..(2) Decision-making model of loan's portfolio optimization based on constraint of the yield of VaR is set up. Taking the loan's yields as the profit of financial asset, taking the volatility of loan's yields as reflection of loan's risk, under the constrain of Value at Risk (VaR), based on the solution of quadric programming, a decision-making model of loan-risk portfolio optimization is set up with the minimum risk within the feasible range of definite portfolio yield. .There are three characteristics of the model: Using yield rate of maximum loss but yield amount to reflect VaR, so it becomes convenient for the decision-making analysis. Taking risk correlation into account, controls risk limitation with the maximum loss on yield rate of VaR, so the ability for risk tolerance of commercial bank is reflected by loan's distribution or allocation. If given the objective in the feasible range, and given the yield ratio as decision-maker expected, the loan's portfolio to the minimum risk can always be found. The efficient boundary which is given by this model provides a scientific method for the decision-making of the loan's portfolio. (3) A Model of loan's portfolio optimization based on the fuzzy synthetic evaluation of the earnings is set up.Analyzing the characters and defects of the existed models of loan's portfolio optimization, this research is subject to such constraints of fuzzy synthetic evaluation while aims at minimizing portfolio risks , then establishes a fuzzy optimization model of loan's portfolio. under the condition that the fuzzy synthetic evaluation aim and evaluation matrix of the loan earnings rate are known, we can get the loan's proportion by means of quadratic programming, thus solve the decision-making problem of loan's portfolio. With further practical analysis, it illuminates the model is reasonable and feasible..The characteristics lies on two aspects: Firstly, the fuzzy synthetic evaluation method is brought to the model of loan's portfolio optimization. Looking on the loan's proportion as the proportion fuzzy set, we can get the loan's proportion reversely in the quadratic programming model if the fuzzy synthetic evalu

建立基于项目风险控制的贷款组合优化原理与模型，使信贷资产的分配反映项目组合的不确定性。建立全部贷款综合风险度控制与优化原理与方法，使贷款增量与存量风险同时得到控.制。建立基于流动性风险控制的贷款组合优化原理与模型，使信贷资产的分配反映资产与负.债结构的匹配与对称。本研究将为银行创建贷款组合优化决策理论与决策方法..