ESTIMATING THE MARKET MODEL ROBUSTLY
针对市场模型参数估计中普通最小二乘法假设不成立的问题，提出一种稳健估计方法，能自动应对非正态性，并用证券样本对比两种方法，发现结果差异显著。
There is now considerable evidence in the literature that the ordinary least squares assumptions fail to hold when estimating the market model parameters. This paper describes a robust estimation procedure which provides automatic protection against departures from normality. The market model parameters are then estimated for a sample of securities using both the least squares method and the robust procedure. Analysis shows that the results under the two procedures may differ considerably.