The Study Of Realized Volatility Based On High Frequency Financial Data

Forecasting and measuring the the volatility of financial asset returns plays a very important role in many areas of financial economics, including risk management, portfolio management and the pricing of derivatives. The increasing information of high frequency financial asset price data has triggered a large number of scholars to estimate and forecast daily volatility. To understand the financial market, it is valuable and meaningful to study volatility based on high frequency financial data.This paper makes a research on volatility of financial asset returns, provides a novel jump-robust estimator based on high-frequency financial data combing the realized method and nearest neighbor-truncation, which is named third minimum realized volatility(TMinRV). The main work and innovation of this paper are organized as follows:Firstly, I summarize several kinds of typical realized volatilities such as realized volatility, realized bipower variation, realized range-based volatility. We compare them on the theory and empirical study.Secondly, this paper provides a novel estimator of integrated variance by the use of non-parametric method, which is the main innovation if this paper. We present the definition TMinRV, prove the consistency and an asymptotic limit theory in the presence of jumps by a series of lemmas. The stochastic simulation and empirical analysis on the statistical property of TMinRV shows that TMinRV can eliminate the influence of price jumps more effectively and estimate asset return volatility more robustness.Finally, this paper applies the theory neighbor truncation to the general integrated power variation estimation based on TMinRV, and structure general third minimum integrated power variation. Then we study the consistency and asymptotic limit theory of this estimator.