论文部分内容阅读
通过对现有理论文献的梳理,提炼了五个较为典型的关于高阶矩风险产生机制的理论假设。然后基于时变高阶矩建模思想,将这五个假设统一于同一个计量框架,并进行综合地实证检验,以期发掘具有“占优”作用的理论解释。以沪深股市收益率为样本研究发现,在这五个假设中,仅“波动率反馈”效应和“利空信息揭示”效应获得显著的实证支持。进一步分析指出,这两种效应会同时引起偏度和峰度风险,因而是高阶矩风险的主要产生机制。这一结果可为后续研究发展一个统一的理论框架提供实证参考。
Through the combing of the existing theoretical documents, five typical hypotheses about the risk generation mechanism of higher moments are extracted. Then based on the idea of time-varying and high-order moments modeling, these five hypotheses are unified in the same metrological framework and a comprehensive empirical test is carried out in order to explore the theoretical explanation of the “dominant” effect. Taking the returns of Shanghai and Shenzhen stock markets as a sample, we find that among the five hypotheses, only “volatility feedback” effect and “negative information revealing” effect obtain significant empirical support. Further analysis pointed out that these two effects will lead to the risk of skewness and kurtosis at the same time, which is the main mechanism of the higher moments risk. This result can provide an empirical reference for the development of a unified theoretical framework for follow-up research.