摘要
本文采用均值-方差张成的方法研究B股相对于A股的市场投资价值。首先,通过MonteCarlo模拟研究了均值-方差张成检验的小样本性质,发现模型的GMM-Wald检验存在显著的小样本偏倚,故采用基于残差再抽样的Block-Bootstrap方法模拟Wald统计量的分布以克服小样本偏倚的影响;接着分别对A、B股构造规模资产组合作为其资产的代理变量,利用模拟的Wald统计量进行实证研究,结果发现:B股未向国内居民开放前,相对于A股具有投资价值,向国内居民开放后,相对于A股不再具有投资价值。文章最后对B股投资价值的这种变化作出了经济解释,并提出相应的政策建议。
This paper adopts the method of mean-variance spanning to study the market investment value of the B-share relative to the A-share.Firstly,we study the finite sample properties of the mean-variance spanning test by the Monte Carlo simulation,and find that there is remarkable finite sample bias for the model's GMM-Wald test,so we apply the Block-Bootstrap method based on the residual resample to simulate the distribution of the Wald test to overcome this bias.Next,we construct the size portfolio for the A,B share respectively to stand proxy for their assets,using the simulated Wald statistics the empirical research reveals that the B-share has more investment value than the A-share before the B-share open to the domestic resident,but this is not the case after the B-share open to the domestic resident,finally,we make an economic explanation for the change of the B-share's market investment value and put forward the relevant policy suggestion.
出处
《数理统计与管理》
CSSCI
北大核心
2010年第5期899-908,共10页
Journal of Applied Statistics and Management
基金
广东省自然科学基金博士科研启动项目(8451063101000726)
国家自然科学基金青年项目(10801137)
教育部人文社会科学项目资助(06JC790016)