-
首页
-
教学项目展开 / 收起
sidenav background教学项目
sidenav header background[11月15日]计量、金融和大数据分析workshop
发布日期：2019-11-12 09:21 来源：
【题目】： The Weekly Rhythm of Volatility
【主讲人】：梁方
北京大学国家发展研究院 博士研究生
【时间】：2019年11月15日（周五）10:00-11:30
【地点】：北大经济学院107室
【摘要】：We study weekly patterns in S&P 500 volatility over 21.5 years, from early 1996 to mid 2017, using high-frequency data. About 29.4% of daily variance occurs during the hours the market is closed, but this percentage is time-varying, gradually increasing from about 21% to nearly 40%. Across weekdays, there is substantial variation in overnight variance. The variance of “overnight” returns from Friday to Monday, Thursday to Friday, is nearly 50% higher than other weekdays. We find large and significant announcement effect of the employment data on overnight variances. The close-to-open variances before the announcement days are systematically higher than other days. For open-to-close volatility, the patterns are different. The variances of open-to-close returns in the middle of the week are significantly higher. We propose a bisected realized GARCH model and include categorical variables to study the weekly rhythm of volatility. Our out-of-sample results show that there are substantial gains from modeling overnight returns and open-to-close returns as a bivariate process, rather than simply modeling daily returns.
【主讲人简介】：
梁方，北京大学国家发展研究院金融学博士生
2018至2019年访问美国北卡罗来纳大学教堂山分校经济系
她的研究领域包括金融计量、实证金融、经济预测等
国家发展研究院官方微信
Copyright© 1994-2012 北京大学 国家发展研究院 版权所有, 京ICP备05065075号-1
保留所有权利，不经允许请勿挪用
-
首页
-
教学项目展开 / 收起
sidenav background教学项目
sidenav header background[11月15日]计量、金融和大数据分析workshop
发布日期：2019-11-12 09:21 来源：
【题目】： The Weekly Rhythm of Volatility
【主讲人】：梁方
北京大学国家发展研究院 博士研究生
【时间】：2019年11月15日（周五）10:00-11:30
【地点】：北大经济学院107室
【摘要】：We study weekly patterns in S&P 500 volatility over 21.5 years, from early 1996 to mid 2017, using high-frequency data. About 29.4% of daily variance occurs during the hours the market is closed, but this percentage is time-varying, gradually increasing from about 21% to nearly 40%. Across weekdays, there is substantial variation in overnight variance. The variance of “overnight” returns from Friday to Monday, Thursday to Friday, is nearly 50% higher than other weekdays. We find large and significant announcement effect of the employment data on overnight variances. The close-to-open variances before the announcement days are systematically higher than other days. For open-to-close volatility, the patterns are different. The variances of open-to-close returns in the middle of the week are significantly higher. We propose a bisected realized GARCH model and include categorical variables to study the weekly rhythm of volatility. Our out-of-sample results show that there are substantial gains from modeling overnight returns and open-to-close returns as a bivariate process, rather than simply modeling daily returns.
【主讲人简介】：
梁方，北京大学国家发展研究院金融学博士生
2018至2019年访问美国北卡罗来纳大学教堂山分校经济系
她的研究领域包括金融计量、实证金融、经济预测等
访问学者申请
EMAIL
北京大学
ENGLISH
