摘要
本文认为,对金融资产尖峰厚尾特性的描述,分形分布比正态分布更合适。我国股票市场收益率分布的特征指数不是正态分布,具有明显的尖峰态,我国股票市场具有较大的波动性;上海和深圳的两个股票市场收益率呈负偏,股票收益率分布右偏,呈现右厚尾特征;当置信水平为99%时,用分形分布拟合经验分布得到的风险系数高估了风险,用正态分布低估了风险,且分形分布的绝对偏差大于正态分布的绝对偏差,而当置信水平为90%时,分形分布对经验分布的风险系数拟合得非常好,正态分布对经验分布高估了风险,且绝对偏差比较大。
The distributional form of return of asset is the key factor in study of asset risk. Due to the financial asset in reality has the characteristic of sharp peak and heavy tail, applying the fractal distribution to research the risk characteristics of financial asset is more significant. This paper researches into the risk characteristics of eight stocks in Hu-shen stock market by applying fractal distribution. It would be found that fractal distribution can fit the rate of return of financial asset better through Fourier transformation and Kolrnogorov criterion, but the normal distribution can＇t. On the other hand, through the method of Monte Carlo Simulations, It would be found that the in a certain confident level. Compared with experienced distribution, when the confident level is lower, the effect of fractal distribution is better than that of the normal distribution. On the contrary, the effect of normal distribution is better than that of fractal distribution.
出处
《中国流通经济》
CSSCI
北大核心
2007年第3期59-61,共3页
China Business and Market
关键词
分形分布
风险价值
特征函数
fractal distribution： VaR
characteristic function
