A Study Of Leverage Effect On High Frequency Financial Data

In recent years,with the deepening of scholars’research on High-Frequency Fi-nancial data,the relationship between variables of high-frequency financial data has gradually become a new topic in this field.The leverage effect describes the negative correlation between asset price returns and volatility changes,the study will help people understand the mechanism of price changes in high-frequency areas.This paper quantifies the leverage effect,for the price process and volatility process of logarithmic assets satisfying the Ito semi-martingale model,we decompose the lever-age effect into two parts:continuous and discontinuous,and construct statistics to estimate them,further study the asymptotic properties of the estimators,and verify the good performance of the estimators with actual data.The main work of this paper lies in:1.The continuous partial leverage effect(CLE)is studied.We Combine prob-ability limit theory and stochastic process theory,and the idea of near window and downward truncation,Monte Carlo simulation method is used to explain the large sample properties of the(CLE)estimator,and the good performance of the estimator is verified by actual data.2.The discontinuous partial leverage effect(DLE)is studied.The large sample properties of the(DLE)estimator are discussed by using the idea of proximity win-dow,upward truncation and probability limit theory.These properties are further verified by numerical simulation and empirical analysis.