城投债违约风险测度与预警——基于KMV和BP模型

A Pre-warning System for Default Risk of Urban Investment Bond-Based on the KMV-BP Model

结合KMV模型和BP神经网络模型构建城投债违约风险预警模型,以区域经济实力、发行主体财务状况、发行特征3个维度22个指标搭建城投债违约风险评价体系,识别提炼我国城投债违约风险的关键影响因素,考察分析各维度特征以及异质性表现。通过对6675只债券,14万条以上总样本进行实证检验后发现,债券发行主体财务状况是在评价同一地区城投债违约风险时最具参考价值的维度;同省份不同城市的城投债违约风险水平因区域风险冲击而趋同;城投债因所处地理区位、发行主体所属行业不同,其风险表现具有显著差异。研究结论有助于投资者建立风险预警机制,有效防范城投债的违约风险。

Based on the KMV model and neural network model,an pre-warning system is constructed for urban investment bond default risk with 22 indicators in 3 dimensions:regional economic strength,financial status of the issuing entity and issuance characteristics,to identify the key influencing factors of urban investment bond default risk in China,and to examine and refine the characteristics of each dimension and the heterogeneity of urban investment bonds in each province.An empirical examination of 6,675 bonds with a total sample of over 140,000 notes reveals that the financial status of bond issuers is the most valuable dimension in evaluating the default risk in the same region;the default risk level in different cities of the same province converges due to regional risk shocks;and the performance of urban investment bonds in different geographical locations and industries of the issuers differs significantly.The findings of the study will help investors to establish a risk warning mechanism to effectively prevent the risk of default on urban investment bonds.