摘要
针对分位回归模型参数的不确定性问题,构建基于贝叶斯分位数回归的联动风险价值模型,据此研究金融风险的相依性问题。通过模型的统计结构分析,选择参数先验分布,设计贝叶斯MCMC算法估计参数。并利用机构总资产收益率进行实证分析。研究结果表明:贝叶斯分位回归模型可以有效地描绘极端风险下的相依性,金融业的风险相依性大于实体行业。
In order to research financial risk interdenpdence, a conditional value at risk model based on Bayesian quantile regression for the uncertainty risks of quantile regression model parameters is constructed. Based on analysis of statistic of model and the selection of parameters prior,the Bayesian MCMC is utilized to estimate model parameters. The empricial research applies the growth rate of institutions' market valued total asset. The research results show that the interdependece of the risk can be accurately estimated using bayesian quantile regression. The interdependece of the risk in the financial sector is stronger than that in the real economy.
出处
《中国管理科学》
CSSCI
北大核心
2016年第S1期480-488,共9页
Chinese Journal of Management Science
基金
国家自然科学基金创新群体资助项目(71521061)
国家自然科学基金重点资助项目(71431008)
关键词
风险相依性
联动VaR
分位数回归
贝叶斯分析
MCMC算法
状态变量
risk interdependence
conditional value at risk
quantile regression
Bayesion analysis
MCMC algorithm
state variables
