动态贝叶斯信用评级的宏观经济冲击模型  

Macroeconomic Shock Model for Dynamic Bayesian Credit Rating

作　　者：吴建华[1] 张颖[1] 原雪梅 WU Jian-hua;ZHANG Ying;YUAN Xue-mei(School of Mathematical Sciences,University of Jinan,Jinan 250022,China;Institute of Finance,University of Jinan,Jinan 250002,China)

机构地区：[1]济南大学数学科学学院,山东济南250022 [2]济南大学金融研究院,山东济南250002

出　　处：《数理统计与管理》2022年第6期969-981,共13页Journal of Applied Statistics and Management

基　　金：国家社会科学基金一般项目(17BJY184);国家自然基科学金(11701214);山东省自然科学基金项目(ZR2018LG001);山东省高等学校人文社科计划一般项目(J17RA103)。

摘　　要：本文考虑了宏观经济金融指标和潜在的系统性冲击,提出了动态评价信用质量的贝叶斯模型,以处理在债券信用风险量化中存在的样本量较小和违约事件稀疏问题。模型的动态特征可以解释相同信用等级在不同时期的违约概率之间的相关性,模型中宏观协变量可以解释宏观经济状况对债券市场信用质量的影响,潜在冲击项可以解释在给定时期内的不同等级的转移概率之间的相依性。利用2009年到2018年中国债券市场上相关债券的财务、非财务和历史评级数据对模型进行了实证检验。结果表明模型能够有效的估计低违约组合的非零违约概率,可以解释宏观经济变量和潜在冲击对不同等级债券的影响力,以及信用等级转移中存在的相依性。模型可以为监管部门、评级机构和债券投资者提供一个新的模型选择思路。This paper considering the macro-economic and financial indicators and the potential systemic impact,a Bayesian model for dynamic evaluation of credit quality is proposed to deal with the problems of small sample size and sparse default events in the quantification of bond credit risk.The dynamic characteristics of the model can explain the correlation between default probabilities of the same credit rating in different periods,the macro covariables in the model can explain the impact of macroeconomic conditions on the credit quality of the bond market,and the potential shock term can explain the dependency between the transfer probabilities of different ratings in a given period.The financial,nonfinancial and historical rating data of relevant bonds in China ’s bond market from 2009 to 2018 are used to test the model empirically.The results show that the model can effectively estimate the non-zero default probability of low-default portfolios,and can explain the influence of macroeconomic variables and potential shocks on bonds of different grades,as well as the dependence of credit grade transfer.The model can provide a new model choice for regulators,rating agencies and bond investors.

关 键 词：宏观经济冲击 债券市场 评级转移 贝叶斯推断 

分 类 号：F830.91[经济管理—金融学] O212[理学—概率论与数理统计]