具有未知跳跃股票价格异方差性的 Bootstrap模拟检验

采用Bootstrap模拟方法检验股票价格异方差性．利用随机分析方法获得波动率的一个渐进估计，这种估计能够消除股票价格中可能存在的随机跳跃的影响，并且无需确定波动率函数的具体形式、建立了一个关于波动率为常数的原假设的统计量，用Bootstrap模拟方法可以获得该统计量的分布．用此检验方法对中国沪深两市指数进行检验，获得了两市中指数呈异方差性的结论．这种检验方法对认识市场微观结构、资产定价、投资和风险管理都具有积极作用．

Bootstrap test on the heteroskedasticity of stock price with unknown jumps

This paper tests on the heteroskedasticidy of stock prices with bootstrap method. With stochastic analysis method, an asymptotic estimation of volatility is obtained, which can rule out the effect of unknown random jumps and needs no specification of volatility functions. Then a statistics with null hypothesis of unvarying volatility is built, whose distribution is obtained by bootstrap simulation. Tests on the heterosdedasticity of both Shanghai and Shenzhen security indexes are carried out and the results show that heteroskedasticity of both indexes exists. This method can be helpful for better understanding the microstructure of financial market, capital pricing, investment and risk management.