危机传染效应的识别与度量——基于改进MIS-DCC的分析  

Identification and measurement of contagion effects of the crises: Based on improved Markov independent switching dynamic conditional correlation model analysis

作　　者：苏海军[1,2] 欧阳红兵[2] 

机构地区：[1]招商银行博士后工作站,深圳518067 [2]华中科技大学经济学院,武汉430074

出　　处：《管理科学学报》2013年第8期20-30,共11页Journal of Management Sciences in China

基　　金：国家自然科学基金资助项目(70971051)

摘　　要：论文对MS-DCC和IS-DCC提出改进,构建了Markov独立转换的动态条件相关(MISDCC)分析框架,使其兼具算法估计和传染分析的优越性,进而考察美国次贷危机、欧洲主权债务危机在主要证券市场间的传染效应.改进模型内生地刻画出了传染的不同区间,避免了将考察期武断地分割为危机前、后的误区.阐明了次贷危机、欧洲主权债务危机在当前开放市场间的传染是系统性的,应对危机需要各国的协调配合.证据显示美国及其他有关国家贻误了应对次贷危机的时机.危机期市场中的信息更加复杂,以至于市场在各机制间较为频繁地转换;市场的条件和无条件相关可以借助机制变量进行更直观的描述.Based on Markov switching dynamic conditional correlation （MS-DCC） and independent switching DCC （IS-DCC） models, we construct a Markov independent advantage to be estimated and analyzed accurately. Then we switching DCC （MIS-DCC） model which has the investigate the contagion effects among five major stock markets from the American subprime mortgage crisis to European sovereign debt crisis. We find that, firstly, our model endogenously characterizes different intervals of the contagion, which effectively avoids the shortcoming that divides the whole sample into pre-crisis and post-crisis period arbitrarily. Secondly, we find it is necessary for countries to cooperate with each other to respond to crises since American subprime mortgage crisis and European sovereign crisis evolve into a systemic risk. Thirdly, there is the evidence that America and other relevant countries miss the time to respond to the American subprime mortgage crisis. Fourthly, we find more complex information in period of the crisis, which results to frequently switch between different re- gimes of the market. Finally, both conditional and unconditional correlations can be described intuitively by regime variables.

关 键 词：关联结构 MARKOV 独立转换 传染效应 

分 类 号：F830.91[经济管理—金融学]