主 题: Trading Information, Price Discreteness and Volatility Estimation
内容简介: This paper studies the problem of volatility estimation using high-frequency financial data under a general Roll (1984) type microstructure model where asset prices are partially observed with discreteness. The problem is challenging in that efficient prices are perturbed by both random microstructure noises and a nonlinear rounding effect. We establish the asymptotics of a rounding effect on the recovery of trading-information-based microstructure noises under a small error paradigm. Armed with these results, we find that a particle filter coupled with nonlinear least squares method can effectively remove biases caused by the mixed-type microstructure noises in volatility estimates. Simulation and empirical results demonstrate the superiority of our method over existing popular high-frequency volatility estimation methods under comparison.
报告人: 张志远 副教授 博士
时 间: 2018-07-08 09:30
地 点: 竞慧东楼302
举办单位: 统计与数学学院 统计科学与大数据研究院