异质金融市场驱动的已实现波动率计量模型

本文根据金融高频数据的典型特征和异质市场假说,首次提出了异质市场驱动因素的分层结构,并借此构建了已实现波动率的HAR-L-M计量模型。模拟分析显示出该模型的合理性和优越性,样本内、外预测都说明HAR-L-M模型优于现有已实现波动率HAR模型和ARFIMA模型。最后的实证分析结果显示,中国市场的异质程度要强于美国证券市场,同时个股更容易受多种异质驱动因素的影响,个股稳定性要比股指差。

A New Realized Volatility Model Driven by Heterogeneous Finance Market

Based on the stylized facts of volatility observed in financial high-frequency data,and heterogeneous market hypothesis,we evolve a hierarchical structure of heterogeneous market-driven factors,and produce a new realized volatility model,termed as HAR-L-M.The simulation shows its rationality and superiority,and this model shows remarkably better in and out of sample forecasting performance than other realized volatility models,such as HAR and ARFIMA etc.At last,empirical analysis shows that degree of hetero...