The Study of Jump Behavior of Chinese Asset Prices:Evidence from High frequency Data Sets

Current studies on financial markets' asset prices' jump behavior usually use parameter estimation based on models which set various parameters for jumps in analysis of asset prices.This paper uses intra-day high frequency data to model volatility with realized volatility model and the continuous part of volatility with realized second power volatility model,which is the basis of the non-parametric estimation of asset prices' jump behavior.This paper also separates the days on which prices jump,identify the size and direction of jumps.Empirical results show that jumps happen heavily on days which are very volatile,and the size and durations of jumps also present clustering phenomenon.The jump densities of five representative assets are all left-skewed,which means the down-side jumps happen more often than up-side jumps.