Research On The Application Of GARCH - VaR Model In China 's Stock Market Under Noise

With the development of behavioral finance theory, the noise trading has received widespread attention by many scholars. They have already been studying the influence of noise trading on capital market. China’s stock market started relatively late, therefore, there are still a lot of problems in China’s stock market, for example the imperfect system of stock market, the investors of lacking professional knowledge and so on. Especially, the performance of the noise trading in China’s capital market is prominent. Under this background, this paper mainly research the influence of noise trading on volatility and risk measurement in China’s stock market.Firstly, this paper introduces the related concepts of noise trading and analyzes the influence of noise trading on volatility of risky assets in theory. Then based on the actual situation of the capital market, we use the method of factor analysis to measure the level of the noise trading.Secondly, we introduce how to measure the volatility and risk based on the nonparametric GARCH model, and combining with the relevant theory of noise trading we establish the nonparametric GARCH- VaR model in the case of noise trading.Finally, based on the related data of January 5, 2009 to March 31, 2014 in Shanghai stock market, we measure the volatility and VaR in the Shanghai stock market by nonparametric GARCH- VaR model in two different situations, which one is in the ordinary situation and the other is in the case of noise trading. Then, we compare the two groups of the results and conclude the following conclusions: 1、In the stock market, the noise trading has significant influence on the volatility of the market, and the noise trading in the market would increase the market volatility. 2、Compared with the two groups of results in the two different situations, the accuracy of the two results is at the same level, however, the result in the case of noise trading express less losses. Therefore, based on the two points, we can conclude that the nonparametric GARCH model in the case of noise trading have better prediction effect.