基于递归均值调整非平稳波动率资产价格泡沫检验量分布与实证研究

为提高非平稳波动率资产价格泡沫检验量的功效,首先使用递归均值调整方法构建PWY检验量,并讨论在特定数据生成过程下的检验量分布; 其次利用Wild Bootstrap方法构建联合检验量; 最后进行蒙特卡罗模拟研究和实证研究.理论研究表明:检验量在大样本下均收敛于维纳过程的泛函,含有非平稳波动率信息,且与已有检验量分布不同.模拟研究显示:递归均值调整PWY检验量功效在弱激增数据生成过程中优势明显,经典PWY检验量功效在中度激增数据生成过程中有微弱优势,联合PWY检验量功效位于2者之间.实证研究显示:递归均值调整PWY检验量的结论更符合实际.

The Distribution and Empirical Research of Asset Price Bubble Test with Non-Stationary Volatility Based on Recursive Mean Adjustment

Firstly,to improve the power of asset price bubble test with non-stationary volatility,the PWY statistic is constructed by using the recursive mean adjustment method,and the distribution of the statistic is discussed under the specific data generation process.Secondly,the Wild Bootstrap method is adopted to construct the union statistic.Finally,Monte Carlo simulation and empirical research are carried out.The theoretical study shows that the distributions of statistic converge to the function of Wiener process under large samples which is different from that of the existing one,and contain non-stationary volatility information.The simulation study shows that the power for the recursive mean adjusted PWY statistic is higher than that of classical PWY statistic when the explosive parameter is relatively weak,and the conclusion is opposite for moderate parameter.The power of the union PWY statistic is in between.The empirical research shows that the conclusion of the recursive mean adjusted PWY statistic is more consistent with the reality.