The Impacts Of Measurement Errors On The Volatility Of Chinese Stock Market

As a measure of the risk of financial assets,volatility plays an important role in many financial topics.Therefore,accurately measuring and forecasting financial volatility is of crucial importance.Thanks to the ever-developing technology,it's easier to acquire high-frequency data than before,which results in a shift from commonly used GARCH models to more recent Realized Volatility models.The HAR-RV model based on Realized Variance is the benchmark in this area.As the sampling frequency goes to zero,realized volatility is consistent for the true latent volatility,however in any given finite sample it is,of course,subject to measurement error,standard approaches for dealing with errors-in-variables problem treat the variance of the measurement error as constant through time.In contrast,my study explicitly takes into account the temporal variation in the errors when modeling and gives a deep investigation on impacts of measurement errors on the volatility of Chinese stock market.Empirical results show that during the years 2007 to 2008,and 2014 to 2015,the Chinese stock market exhibits a high volatility period,measurement errors seem to cluster in these periods too.Introducing measurement errors into the HAR-RV model forms HARQ model,HARQ model shows both superior in-sample fit and out-of-sample forecast.The data of Chinese stock market and empirical results suggest there should be adjustment for measurement errors in longer horizon RVs,innovatively,by introducing an extra autoregressive parameter adjustment term in the HARQ model which forms HARQ-A model,it keeps the simplicity and facilitate processing initial data,empirical results show the new HARQ-A model outperform basic HAR-RV and HARQ models in both in-sample fit and out-of-sample forecast for Chinese stock market,thus the HARQ-A model helps to forecast volatility of Chinese stock market more accurately.