|
|
|
Statistics/Probability Seminar |
Spring 2007 |
DATE: Friday, April 20, 2007 |
TIME: 12:30 - 1:30pm |
ROOM: MAP 213 |
Multiscale Jump and Volatility Analysis for High-Frequency Financial Data |
Professor Yazhen Wang |
Department of Statistics, University of Connecticut |
|
ABSTRACT:
Volatilities of asset returns are pivotal for many issues in
financial economics. The availability of high frequency intraday
data allows us to accurately estimate stock volatility.
Because asset prices often have jumps, and high-frequency data are
contaminated with market microstructure noise, common approach of
estimating integrated volatility is to sample from available data
and use the obtained subsample (a fraction of the original data) to
compute realized volatility and realized bipower variation. In this
paper we propose wavelet based multiscale methods to perform jump
analysis in price processes, and estimate integrated volatility and,
if jump(s) occurs, recover jump variation. We establish asymptotic
theory for the methods and study their efficiency for finite sample.
This is a joint work with Jianqing Fan.
|
|
|