Study On Measurement Of Volatility With Microstructure Noises And Jumps

Financial volatility plays an important role in the perfection of in modern financial theory and practical application. Arbitrage combination, asset pricing, risk management and monetary policy cannot exist without this key variable. In recent years, the popularity of electronic trading and information storage technology provides convenience for the development of high frequency data. High-frequency data can capture the market information quickly and effectively without distortion of information, thus can reflect the real condition of the market.But under the high-frequency environment, market microstructure noise and jumps always exist in the financial market. This will make the volatility biased and inconsistent estimates. A large number of domestic and foreign scholars study on asset price volatility based on the influence of market microstructure noise or jumps. In real financial market, microstructural noise may appear along with microstructural jumps at the same period of time.Market microstructure noise and instantaneous price jumps have little influence on volatility measurement when the sampling interval is large, but big influence when sampling interval is small. It is a hot topic in quantitative economics to research in financial volatility under the influence of market microstructure noise and price jumps.More complicated than economic assumptions, the securities market has an intricate rule of operation. In certain cases, market microstructure noise and price jumps are presented collaboratively, however, few researchers focus on both factors in the literature. To bridge the gap in related academic literature, the paper focuses on the efficient measurement of asset volatility of high-frequency data.In this paper, we measured return volatility which takes microstructure noise and jumps into account by different methods as follows with constant and stochastic simulations: modulated realized volatility, bipower-type statistics, jump-robust volatility estimation using nearest neighbor truncation(MedRV ? MinRV) and pre-averaging realized volatility. Each method is applicable to different markets and different time periods. Which of the improved methods can better measure the volatility of stock index futures needs to be further explored. Firstly we compared the effectiveness of measurement of the four estimators using simulation methods. Toavoid the influence of a small number of singular points on the volatility model, this paper used different forms of loss function as evaluation criteria to determine which estimation method is more suitable for China's securities market. We applied the significant jump test to study the effects of major international environment change,macroeconomic policy and market events to China's stock market return volatility.