论文部分内容阅读
随机理论认为股票收益率服从正态分布,但大量研究表明,股票收益率等金融时间序列具有“尖峰厚尾”反正态性.极值分布和Laplace分布是一类厚尾分布,本文对沪、深两地股票市场的收益率进行了极值分布(以Gumbel分布最为常用)和Laplace分布拟合,并给出了风险价值(VaR)的估计.实证研究表明,Laplace分布是正态分布的一种改进,而Gumbel分布能较准确地估计巨额盈利(极端事件)的风险.
Stochastic theory holds that the stock returns follow a normal distribution, but a large number of studies show that the financial time series such as stock returns have “inverse peak-tail” inverse states. The extreme distribution and Laplace distribution are a kind of thick tail distribution. In Shanghai and Shenzhen stock markets, the extreme value distribution (the most commonly used Gumbel distribution) and the Laplace distribution are fitted, and the VaR estimation is given. Empirical studies show that the Laplace distribution is a normal distribution , The Gumbel distribution can more accurately estimate the risk of huge profits (extreme events).