本课题系统性的研究基于Realized GARCH框架的波动率和相关性模型的计量理论及其在实证金融、衍生产品定价和风险管理领域的应用。该框架由课题负责人及合作者在HHS(2012, forthcoming at JAE)中首次提出，充分利用高频数据已实现波动率中包含的波动率和相关性的精确信息，在波动率快速变化的市场环境中任具有较好的样本内和样本外预测能力，并保留GARCH模型简单优美的数学结构和易于估计的特性。.理论部分研究模型的基本统计性质和估计量的一致性和渐进正态性，以及各方面模型扩展，比如，杠杆效应和测量误差具有不对称传导效应的Realized EGARCH模型；利用模型比较多个已实现波动率的测量误差；考虑厚尾冲击和调整极端值影响的模型等。实证部分主要研究该框架在如下领域的应用：基于高频数据的期权定价模型；股票预期收益率和风险的关系研究；预测动态Beta系数和指数期货对冲策略研究。

This project systematically studies the theory and applications for the volatility and correlation models based on Realized GARCH framework. This framework was originally developed by the Principal Investigator and his co-authors in HHS(2012,forthcoming at JAE). While keeping the GARCH models' nice mathematical structure and the simplicity of estimation, Realized GARCH framework fully utilizes the more accurate information about the latent volatility contained in realized variance computed from high frequency data, and shows good in-sample and out-of-sample forecasting performance.. In the theoretical part, we mainly study the statistical properties, the consistency and asymptotic normality for its estimator for Realized GARCH, and its variants and extensions in several aspects, such as, models allowing asymmetric inter-temporal transmissions for the leverage effect and measurement error, models with multiple realized variances, models considering the fat-tailed shocks and adjusting the impacts of extreme observations. In the empirical part, we mainly study the applications of Realized GARCH on the following issues: option pricing model using high frequency data; forecasting time-varying Beta and dynamic hedging using index futures.