对CAPM模型检验中独立同分布正态假设重要性的实证分析

金融市场的数据大多不是iid正态的,但是通常针对CAPM的Wald检验和F检验均是基于这一假设的。本文采用一种不需要iid正态假定的基于GMM方法的检验统计量,对同一原假设进行了检验,并且将结果与美国和澳大利亚股市的实证结果进行了比较。结果表明上海股市的确不满足iid正态,而且GMM统计量的检验结果和通常的检验存在差异并有可能影响到判别的结果。

The importance of iid-normal assumption in CAPM tests:an empirical analysis

Most of the financial market data are not iid-normal. However, the Wald test and F test for CAPM are based on this assumption. In this paper, we use a GMM estimator, which doesn't need the idd-normal assumption, to test the same null hypothesis in Shanghai stock market, and compare the results with the empirical results of US and Australia. We find that the Shanghai market does not satisfy the iid-normal assumption, and the difference of the test results between the GMM estimator and ordinary estimators may reverse the discrimination.