盈余信息度量、市场反应与投资者框架依赖偏差分析

A Study on Measures of Earnings Information, Market Reaction and Investors' Framing Dependence Bias

本文以2 0 0 0年9月到2 0 0 3年1 2月沪市A股3 3 8家上市公司为研究对象,应用实证研究方法,检验以四种不同度量方式表示同一盈余信息所产生的“盈余惯性现象”是否存在差异。结果表明:在三因素模型进行风险调整之前,基于四种盈余信息指标的盈余惯性现象都显著地存在。但经过三因素模型风险调整后,基于“意外盈余率”和“标准化意外盈余率”二个指标的盈余惯性现象消失了;基于“意外盈余”和“标准化意外盈余”二个指标的盈余惯性现象仍然存在,所以买入赢家组合、卖出输家组合仍可获得显著的超常收益。显然,这一研究结果并不支持风险定价学派的观点。笔者认为,根据Tversky和Kahneman( 1 981 )提出的“框架依赖偏差”(FramingDependenceBias)理论,四个盈余信息指标所产生的盈余惯性现象的差异表明我国投资者对盈余信息的反应依赖于信息度量的方式。

This paper examines and makes a comparison between four earnings momentums that differ in the four measures of earnings, by sampling 338 A-shares listed on the Shanghai Stock Exchange in the period from September 2000 to December 2003. Using each measure of earnings, we find significant evidence of the earnings momentums. But after controlling for risk using the Fama-French three-factor model, the earning momentums that are based on “unexpected earnings ratio” and “standardized unexpected earnings ratio” disappear. However, the earnings momentums based on “unexpected earnings” and “standardized unexpected earnings” are still robust after the control for risk, which means the strategies that buy and sell stocks based on these two earnings measures can generate significantly positive abnormal returns. These empirical results constitute a rejection of the Risk Premium Hypothesis. The authors suggest that these results can be explained by “Framing Dependence Bias” of Tversky and Kahneman (1981). The Chinese investors' reactions to the earnings information vary while different measures of earnings information are used, which results in the difference of four earnings momentums.