This paper uses ARCH models to analyze Shanghai stock market index returns from 19 December 1990 to 30 November 2006, and compares the results of ARCH models with normal distribution and t distribution, finding that the results are very different. The conclusion is that the reasonable results of ARCH models are determined by selecting appropriate distribution of the residuals. T distribution is better suited to fitting the distribution of the residuals than normal distribution. The ARCH effect exists in Shanghai stock market. The leverage effect is significant, and the impact of negative news in the stock market is larger than positive news. Fluctuations caused by impact will be permanent, and can not be eliminated in a short time. The phenomenon of high returns with high risks does not exist in the market.
Published in:
Management Science and Engineering, 2007. ICMSE 2007. International Conference on
Date of Conference: 20-22 Aug. 2007