论文部分内容阅读
利用日内高频数据计算的已实现波动率较好度量了金融资产的风险,因此对其预测模型的研究具有重要意义。考虑到指数成分股的联跳可能蕴含指数跳跃所未能反映的信息,提出运用非参数方法识别指数成分股的联跳,采用自回归条件风险模型估计成分股联跳强度,并将其引入指数的已实现波动率异质自回归(HAR-RV-CJ)模型中,分析模型预测性能的改进。进一步的,考虑到宏观信息公告的发布可能对股市产生整体性影响,相应影响成分股联跳的几率;因此,在成分股联跳的自回归条件风险模型中引入居民消费价格指数、国内生产总值、贸易差额等宏观信息公告变量,并分析对联跳强度估计以及指数已实现波动率预测的影响。采用2011年1月4日至2013年7月11日沪深300指数及其成分股高频数据的实证表明,指数成分股联跳与指数跳跃具有不同的特征;用成分股联跳强度代替HAR-RV-CJ模型中的跳跃构建的HAR-RV-CI模型,较原始的HAR-RV-CJ模型,以及同时考虑指数跳跃与成分股联跳强度的HAR-RV-CJI模型,具有明显较优的样本内拟合与样本外预测性能。引入宏观信息公告变量可以改进联跳强度自回归条件风险模型的拟合效果,并提高指数已实现波动率模型的样本内拟合能力,但对于指数已实现波动率的样本外预测性能并无明显的帮助。
The realized volatility calculated by intra-day high-frequency data better measures the risk of financial assets, so the research on its forecasting model is of great significance. Considering that the jumps of index constituents may contain information that the jumps of the indices fail to reflect, this paper proposes the jumps of non-parametric methods to identify the constituents of the index stocks, and uses the autoregressive conditional risk model to estimate the jump strength of constituent stocks, and introduces them into the index In the realized volatility heterogeneous autoregressive (HAR-RV-CJ) model, the improvement of the model predictive performance is analyzed. Further, taking into account the release of the macro-information announcement may have a holistic impact on the stock market, which in turn will affect the probability of the constituent stocks jumping together; therefore, introducing the consumer price index, the gross domestic product Value, trade balance and other macro-information bulletin variables, and analyze the impact on the estimation of the jump strength and the predicted volatility of the index. The empirical analysis of the HF CSI between the CSI 300 Index and its constituent stocks from January 4, 2011 to July 11, 2013 shows that the index constituent stocks and the index hops have different characteristics; the hop strength of the constituent stocks instead of the HAR The HAR-RV-CI model constructed by hopping in -RV-CJ model is better than the original HAR-RV-CJ model and the HAR-RV-CJI model considering exponential jump and component hop strength. The intra-sample fitting and out-of-sample prediction performance. The introduction of macroscopic information bulletin variables can improve the fitting effect of the jump-risk autoregressive conditional risk model and improve the in-sample fitting ability of the index volatility model. However, the out-of-sample prediction performance of the index volatility has not obvious s help.