摘要
考虑到资产组合有效子集的渐近检验方法通常只适用于大样本情形,为了提高在有限样本和小样本情形下有效子集判定的准确性和稳定性,通过分析资产组合有效子集新的判定条件,提出了一种基于Block-Bootstrap随机模拟的有效子集精确检验方法,同时对几类一致性协方差阵估计下的稳健检验方法进行了检验功效和检验水平的比较,模拟结果表明基于Block-Bootstrap随机模拟的精确检验方法更为可靠,实证结果也表明该方法得到的有效子集更为稳定.
Given the applicable limitation of the asymptotic test of efficient subset of portfolio to the largesample cases, by analyzing the equivalent conditions for determining efficient subset of portfolio, a newtest method based on block-bootstrap simulation was proposed to improve the stability and accuracy oftesting in the small or finite sample cases. The power and size of test were also compared with several testsbased heteroskedasticity-consistent covariance matrix estimator. The results of simulation and empiricalstudy shown that the test based on block-bootstrap simulation was more robust and reliable.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2014年第7期1648-1661,共14页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(71101095)
广东省自然科学基金(2008276)
