标题: | 风险值衡量:实现变幅的应用 Estimating Value at Risk with Realized Range |
作者: | 陈昱如 周雨田 经营管理研究所 |
关键字: | 实现变幅;日内资料;风险值;极值理论;变幅;波动性;Realized range;Intra-day data;Value at Risk;Extreme value theory;Range;Volatility |
公开日期: | 2008 |
摘要: | 本篇论文将实现变幅(realized range)概念应用在风险值模型中,利用Martens and van Dijk (2007) 所提出的修正误差方法,并使用MEM(Multiplicative Error Model)来预测下一期的波动性,得到实现变幅基础下的风险值模型。此外,本研究也利用常态分配假设下的变异数-共变异数法(variance-covariance method),以及厚尾性质的极值理论 (extreme value theory)两种不同假设的风险值模型来一起做比较。在实证上,以标准普尔500(S&P 500)指数与那斯达克(Nasdaq)指数的高频率资料作为研究对象,进行实现变幅、报酬与变幅基础下的风险值模型在风险值的预测能力比较。实证结果显示,以实现变幅为基础下的风险值模型表现优于其他的风险值模型。 This paper investigates the concept of realized range into the Value-at-Risk estimation. We follow the bias-correction method of Martens and van Dijk (2007) and use MEM model(Multiplicative Error Model)to forecast volatility and VaR estimation. In addition, we apply two different VaR methods to make the comparison: Variance-covariance method and Extreme value theory. In empirical research, we use the intra-day data of S&P 500 and Nasdaq Index to compare the forecast ability of VaR with realized range, daily return and daily range data. The comparing result shows that realized-range-based VaR model performs better than other models. |
URI: | http://140.113.39.130/cdrfb3/record/nctu/#GT079637515 http://hdl.handle.net/11536/43041 |
显示于类别: | Thesis |
文件中的档案:
If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.