A Study On The HAR Model Based On High-frequency Data

With the development and improvement of the financial markets, research and exploration of the financial markets become more and more deeply. The research on volatility is play an important role of the financial market’s study. Asset allocation, hedging and risk management are undoubtedly related to the volatility. With the rapid development of computer technology, the high-frequency trading data get more convenient. The researchers have been paid to become an important attention to th realized volatility which based on high-frequency data, and it become a great direct to the estimate way of volatility.In this paper, by the financial empirical methods, analyzed and summarized the results and experiences of other researchers have studied, incorporated the characteristics of the stock market, we build a HAR model based on the high-frequency data in the CSI300Index to study the heterogeneity of the Chinese stock market. In the model, we consider the influence of the volatility which created by the short-term, medium-term and long-term traders’trading behavior to the the volatility of the whole stock market. Furthermore, we improve the original HAR model by introduce a mechanism of state transition to eliminate the residual heteroscedasticity, which is different from general way that incorporate a GARCH model with the HAR model. This method could improve Improve the interpretation of the role of the model, which has been proved in this paper.Analysis of the empirical results, we believe that the improved HAR model has a better explanation to the volatility of the whole stock market. The empirical results of the improved HAR model in which regard the day, weekly and monthly yield as the amount of state is better than the original HAR model get. We also found that in the situation of day yield, the volatility of whole market is caused by the trading behavior of the short-term, medium-term and long-term traders and the short-term jump caused by the material information or other factors. But in the situation of weekly and monthly, the influence degree of short-term, medium-term, long-term traders’trading behavior to the market volatility is dependent on the market trends. And the short-term market fluctuations of external factors on the market might not be available, which is also decided by the market trends.