基于高频数据的中国股市VaR风险研究

发布时间：2018-06-30 00:44

本文选题：VaR + 高频波动率　；参考：《重庆大学》2013年硕士论文

【摘要】：金融市场的风险度量一直是学术界和风险监管当局关注的重点。传统的风险度量大多数都是基于低频日间数据建立的GARCH类模型或SV类模型。虽然这些模型本身能较好的度量时间序列的波动状况，但股市日内交易频繁，低频数据模型会损失大量的日内重要信息。现有研究表明，传统的GARCH类模型并不能直接用于估计高频波动率。建立有效的高频数据风险度量模型，为金融机构和监管当局的风险监控提供一种有效的理论方法参考和政策建议具有重大意义。本文结合前人对已实现类高频波动率的研究，对已实现波动率RV、已实现双幂次波动率RBV和赋权已实现双幂次波动率WRBV进行比较，针对WRBV具有的长记忆性，建立了ARFIMA-WRBV-VaR模型对中国股市风险进行度量，，并与采用低频日间收益率序列建立的GARCH类模型相比较。实证结果表明：ARFIMA-WRBV-VaR模型比EGARCH-VaR模型估计效果更好。而且，已实现类高频波动率出现了跳跃点、日内U型周期性日历效应和长记忆性特征，这些特征受市场微观结构中的信息不对称和投资者心理等因素影响。进而为风险监控提出了完善信息披露机制和增强投资者素质的政策建议。
[Abstract]:Financial market risk measurement has been the focus of academic and risk regulatory authorities. Most of the traditional risk measures are GARCH models or SV models based on low frequency day data. Although these models themselves can better measure the volatility of time series, the intraday trading of stock market is frequent, and the low-frequency data model will lose a lot of important information. Existing studies show that the traditional GARCH model can not be directly used to estimate high frequency volatility. It is of great significance to establish an effective risk measurement model of high frequency data and to provide an effective theoretical and methodological reference and policy advice for the risk monitoring of financial institutions and regulatory authorities. In this paper, we compare the realized volatility RV, the realized double power volatility RBV and the weighted double power volatility rate WRBV with the previous researches on the realized high frequency volatility, aiming at the long memory property of WRBV. The ARFIMA-WRBV-VaR model is established to measure the Chinese stock market risk and compared with the GARCH model based on the low-frequency daytime yield series. The empirical results show that the ratio ARFIMA-WRBV-VaR model is more effective than EGARCH-VaR model. Moreover, the realized high frequency volatility shows jumping points, intraday U-type periodic calendar effects and long memory characteristics, which are influenced by the information asymmetry in the market microstructure and investor psychology. Then it puts forward some policy suggestions to improve the information disclosure mechanism and enhance the quality of investors for risk monitoring.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F832.51

【参考文献】