报告题目：On the estimation of high-dimensional integrated covariance matrix based on high-frequency data with multiple transactions

报告人：夏宁宁 上海财经大学 副教授

照片：

报告时间：2020年8月27日16:00--17:00

报告平台：腾讯会议：258 751 923

邀请人：李本崇

报告摘要：The phenomenon of multiple transactions at each recording time is a common feature for high-frequency data due to the heavy trading and recording mechanism. In this paper, we consider the estimation of integrated covariance (ICV) matrix of high-dimensional diffusion processes based on multiple high-frequency observations. First, for theoretical purpose, we study the time-variation adjusted realized covariance (TVA) matrix proposed by Zheng and Li (2011). We investigate what effects multiple transactions have on the TVA matrix in terms of its limiting spectral distribution (LSD) when the number of stocks and the observation frequency grow in the same rate. Second, for practical use, we study the limiting spectral behavior of the pre-averaging TVA matrix based on noisy multiple observations, since the observed prices are always contaminated by market microstructure noise. We show that the pre-averaging TVA matrix has several desirable properties: it eliminates the effects of microstructure noise and multiple transactions, it allows for rather general dependence structure in the noise process, both cross-sectional and temporal, and its LSD depends solely on that of the ICV matrix. In particular, compared with existing literatures, the most important property of the pre-averaging TVA matrix is that the aforementioned desirable properties still hold in the presence of asynchronicity, where there are different multiple-transaction numbers for different stocks during each time interval. Third, based on the pre-averaging TVA matrix, we further propose a nonlinear shrinkage estimator of ICV matrix through splitting of the data. We prove that the proposed estimator is asymptotically positive-definite and enjoys a desirable estimation efficiency.

At last, simulation studies support our theoretical results and an analysis of stock market data demonstrates impressive performance of our proposed estimator.

报告人简介：夏宁宁，上海财经大学统计与管理学院副教授、硕士生导师。2013年博士毕业于新加坡国立大学。主要从事高维数据分析，随机矩阵理论，高频数据分析，因子模型领域的研究。已在统计学顶级期刊Annals of Statistics等杂志上发表论文多篇。主持国家自然科学基金2项, 上海市浦江人才计划项目一项。