对于无法获得具体财务信息的银行投资项目,银行需要通过项目宏观风险来推断项目风险进而对投资行为进行决策。以项目的政策风险、政府风险以及行业风险反映项目的宏观风险,本文建立了银行项目宏观风险的评级模型。本文的创新在于,一是通过投影寻踪模型将多个项目风险指标投影为单个综合指标,反映了项目宏观风险要素的非线性组合特征,解决了银行项目宏观风险综合分值的测算问题。二是通过差异序列的有序聚类法对项目综合得分值进行有序聚类,定量测算了项目宏观风险的等级临界值,建立了项目宏观风险的评级模型。
When banks are not able to obtain the financial information of the investment project, they have to use the macro risks of the project to infer project risk and decide whether to invest. This paper selects three risk elements including policy risk, govern- ment risk and industry risk to reflect the macro risks of the project, and establishes rating model of the bank project macro risks. In this paper, the main innovation are: first, this paper uses Projection Pursuit Model to project multiple projects risk indicators for a single composite indicator, reflects non-linear combination characteristics of project macro risk factors, solves the measurement problem of comprehensive scores of project macro risk. Secondly, this paper uses ordered clustering method of different sequences to cluster comprehensive scores of the projects, measure threshold of project macro risks, establish the rating model of project mac- ro risks.