主题：Inference for Volatility under Time Endogeneity
主讲人：崔文昊 北京航空航天大学
主持人：朱海斌 暨南大学
时间：2024年4月12日（周五）下午15:00-16:30
地点：暨南大学石牌校区经济学院大楼（中惠楼）102室
摘要
We propose an endogenous sampling scheme for volatility estimation and inference. The proposed scheme minimizes the asymptotic variance of realized variance, resulting in significant efficiency gains compared to equidistant observations due to its incorporation of additional informational content from the observation times. Furthermore, we develop valid inference procedures for both fixed-k and large-k scenarios for the entire spot variance process. In both settings, our method allows for the construction of uniform confidence bands over a nontrivial time interval while maintaining satisfactory size control. Both inference procedures result in narrower confidence intervals compared to equidistant observations. We also extend our inference procedures to account for discretization error and market microstructure noise. An empirically calibrated simulation study demonstrates the practical reliability of our proposed inference procedures in the presence of pricing errors. Additionally, an empirical application using high-frequency data illustrates the efficiency gains achieved by adopting our method.
主讲人简介
崔文昊，北京航空航天大学经济管理学院经济统计系讲师，2019年于北卡州立大学获得经济学博士学位。主要研究领域为高频金融计量经济学、计量经济学理论和时间序列分析，相关论文发表在Journal of Business & Economic Statistics, Journal of Econometrics, 计量经济学报，担任多个计量经济学期刊的审稿人。主持国家自然科学基金青年项目。
欢迎感兴趣的师生参加！
校对|朱海斌
责编|彭毅
初审|姜云卢
终审发布|何凌云
（来源：暨南大学经济学院微信公众号）
