商业银行发票贷的风险管理及决策支持模型研究

Construction of Risk Management and Decision-making Model of Commercial Bank Invoice Loan

由于中小微企业目前面临着融资难与资金短缺的巨大矛盾,缓解银行和中小微企业之间信息不对称的问题对企业健康稳定成长、对国民经济高质量发展至关重要。以“发票贷”为研究对象,建立能够充分挖掘中小微企业发票信息的指标体系,并采用二元Logistic回归模型对企业的违约概率进行计算和预测。同时,针对贷款定价,在引入RAROC定价模型确定基准利率的基础上,考虑贷款额度带来的风险溢价,分析利率与额度之间的关系。兼顾银行最大化收益和最小化风险的目标,以组合风险价值VaR为约束条件建立多目标规划模型,并采用Python Geatpy库中的遗传算法进行求解。实证研究表明:此模型满足了商业银行对不同目标排序的动态需求,具有科学性和可推广性,可以较好地满足实践需要。

As small and medium-sized enterprises are currently facing the difficulties of financing and shortage of funds,it is crucial for the healthy and stable development of enterprises and high-quality development of the national economy to alleviate the problem of information asymmetry between banks and small and medium-sized enterprises.We take the“invoice loan”for small and medium-sized enterprises as the research object,establish an index system that can fully mine invoice information,and use the Binary Logistic Regression model to comprehensively predict the default probability of enterprises from both quantitative and qualitative perspectives.For loan pricing,we adopt the RAROC pricing model to figure out the benchmark interest rate.Then considering the risk premium brought by the loan amount,we construct the relationship between the interest rate and amount.Taking into account the banks'goals of maximizing returns and minimizing risks,we establish a multi-objective programming model with the constraint condition of the loan portfolio's VaR and use the genetic algorithm in the Python Geatpy library to solve the problem.Empirical research shows that this model can meet the needs of banks for the dynamic change of targets in practice.