金融实证分析中应注意样本容量的确定

Noting the Determination of Sample Size in Empirical Financial Analysis

人们在金融实证分析中,很少关注如何确定样本容量(样本大小)的问题。本文以GARCH模型为例,通过递归估计参数和相应的矩条件,并结合上海股市的具体情况进行分析,结果表明上海股市波动率GARCH模型最优样本容量是三到四年。本文最后从中国金融制度的变化、样本容量对研究结论的影响等方面探讨了样本容量的确定问题。

People rarely pay attention to the determination of sample size in empirical financial analysis. The paper, by GARCH models and through recursive estimation parameters and corresponding moment conditions, empirically analyzes the Shanghai stock market, and reaches the conclusion that the optimized sample size of GARCH model of shanghai stock fluctuation is three to four years. Furthermore, the paper probes into the determination of sample size from the financial system changes of China and the effects of sample size on research conclusion.