摘要
考虑多阶段状态变量的动态信息对违约风险的影响,同时考虑宏观因素和公司个体因素来构建违约预报模型,并且通过在状态变量中包含的行业因素来刻画行业间可能存在的信用传染效应;建立违约风险强度中参数的极大似然估计和渐近性质,进而建立条件违约概率期限结构的极大似然估计.利用极大似然估计及其渐近性质考虑传染效应的显著性检验问题,最后通过模拟研究比较文中所给出的两种估计方法和检验方法的表现.
In this paper we consider the default risk models with muhi-period stochastic covariates, incorporating the dynamics of macroeconomic, firm-specific and industry-specific covariates which were used to characterize the credit contagion between industries. We propose maximum likelihood estimators for parameters of default risk intensities, and then show the maximum likelihood estimator for term structure of conditional corporate default probabilities. Meanwhile, it is easy to construct the significant test for credit contagion effect in multi-period corporate default risk models by the maximum likelihood estimators. Finally, simulation studies are conducted to compare the performance of the working independence estimator to that of the weighted marginal likelihood estimator.
出处
《管理科学学报》
CSSCI
北大核心
2011年第1期1-12,共12页
Journal of Management Sciences in China
基金
国家杰出青年基金资助项目(70825004)
国家自然科学基金重点资助项目(10731010)
国家自然科学基金委创新研究群体科学基金(10721101)
国家973项目子项目(2007CB814902)
上海财经大学"211工程"三期重点学科建设资助项目
上海市重点学科建设资助项目(B803)
关键词
金融危机
违约风险
信用传染
违约预报
financial crisis
default risk
credit contagion
default prediction