基于时间序列极值理论的跳跃风险研究

金融资产的跳跃行为作为对极端事件的刻画,为研究极端事件风险提供了良好工具。基于时间序列下的极值理论,在放松独立同分布假设下,构造了金融资产收益率序列尾部中跳跃动态特征的极值模型。通过对上证综指大跨度、高频度的实证研究,剖析了投资者结构、投资者行为与收益尾部分布之间的相互作用机制,进一步对金融资产收益尾部的跳跃风险进行了有效测度。结果表明,极端跳跃风险的分布特征在频率与尾部方向上呈现很强的不对称状态。

Research on Jump Risk Based on Extreme Value Theory of Time Series

The characterization of jumps of financial assets provided a perfect tool for the study of the risk of extreme events in financial market.Based on extreme value theory of time series and relaxed assumption of independent and identically distributed,we constructed a extreme value model which could capture the dynamic characteristics of the jump in the downside of the financial asset return series.Through the empirical research on stock index of Shanghai,we analyzed the investor structure,the investor behavior and the interaction mechanism between them and downside return series.Furthermore,we measured the jump risk of the downside return series of the financial assets.The results showed that jump risk is asymmetry distributed in the frequency and direction.