基于高频数据的金融波动率建模研究

项目名称： 基于高频数据的金融波动率建模研究

项目编号： No.71201075

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 管理科学与工程

项目作者： 瞿慧

作者单位： 南京大学

项目金额： 22万元

中文摘要： 波动率是现代金融理论中对资产风险的主要度量，它对投资决策与风险管理等金融活动具有重要意义。高频数据的日益可得使我们可以构建"已实现估计量"，将波动率由隐变量转变为可观测的显变量，直接分析、建模；并使得对连续性波动与跳跃性波动的进一步区分成为可能，因此基于高频数据的金融波动率建模研究已成为波动率研究领域的热点。本项目选取我国证券、期货市场的日内高频数据，在现有研究基础上，分别从"已实现估计量"的构建、波动率模型外生变量的选择、波动率模型结构的设计三方面改进创新。通过将每日跳跃性波动进一步合理划分为不同规模的跳跃，有效检测日内多次跳跃，合理纳入"隔夜收益"、宏观经济变量等外生变量，合理引入非线性结构，优化HAR类模型时间尺度结构等方法，改进波动率模型的数据刻画能力与预测性能。本项目将为我国投资者的资产配置与风险管理等金融活动提供有效基础工具，对于政策制度制定者及金融监管当局也具有重大实际意义。

中文关键词： 高频；已实现波动率；跳跃；波动率预测；异质自回归模型

英文摘要： As the quantification of the financial instrument's risk, volatility is of major importance for financial activities such as investment and risk management. The availability of financial high frequency data enables us to construct nonparametric "realized measures", which transfers volatility from hidden to observable variable that could be directly analyzed and modeled, as well as makes it possible to separate continuous-time volatility from jump volatility for further analysis. Thus financial volatility modeling using high frequency data is currently the main stream of volatility research. Using intraday high frequency trade and quote data of China's securities and index futures markets, this project improves the current high frequency volatility modeling research from three aspects: the construction of "realized measures", the choice of exogenous variables, and the design of model structure. More specifically, we improve the data fitting and forecasting capability of high frequency volatility models through reasonably separating daily jump volatility into jumps of different sizes, effectively detecting intraday jumps, appropriately bringing in exogenous variables such as "overnight returns" and macroeconomic variables, incorporating non-linear structures, optimizing the time-scale structure of HAR models, etc.

英文关键词： High-frequency；Realized Volatility；Jumps；Volatility Forecasting；Heterogeneous Autoregressive Model